Vous êtes sur la page 1sur 564

research report

Direct Strength Design of


Cold-Formed Steel
Members with
Perforations
RESEARCH REPORT RP09-1
MARCH 2009

Committee on Specifications
for the Design of Cold-Formed
Steel Structural Members

American Iron and Steel Institute

The material contained herein has been developed by researchers based on their research
findings. The material has also been reviewed by the American Iron and Steel Institute
Committee on Specifications for the Design of Cold-Formed Steel Structural Members. The
Committee acknowledges and is grateful for the contributions of such researchers.
The material herein is for general information only. The information in it should not be
used without first securing competent advice with respect to its suitability for any given
application. The publication of the information is not intended as a representation or warranty
on the part of the American Iron and Steel Institute, or of any other person named herein, that
the information is suitable for any general or particular use or of freedom from infringement of
any patent or patents. Anyone making use of the information assumes all liability arising from
such use.

Copyright 2009 American Iron and Steel Institute

TheJohnsHopkinsUniversity
DepartmentofCivilEngineering
LatrobeHall210
Baltimore,MD21218

ResearchReport

DIRECTSTRENGTHDESIGNOF
COLDFORMEDSTEELMEMBERS
WITHPERFORATIONS

by

CristopherD.Moen
GraduateResearchAssistant

B.W.Schafer
AssociateProfessor

Submittedto:

AmericanIronandSteelInstitute
CommitteeonSpecificationsfortheDesignofColdFormedSteelStructuralMembers

1140ConnecticutAve,Suite705

Washington,DC20036

March2009


Abstract

Coldformedsteel(CFS)structuralmembersarecommonlymanufacturedwithholes

toaccommodateplumbing,electrical,andheatingconduitsinthewallsandceilingsof
buildings.Currentdesignmethodsavailabletoengineersforpredictingthestrengthof
CFS members with holes are prescriptive and limited to specific perforation locations,
spacings,andsizes.TheDirectStrengthMethod(DSM),arelativelynewdesignmethod
forCFSmembersvalidatedformemberswithoutholes,predictstheultimatestrengthof
ageneralCFScolumnorbeamwiththeelasticbucklingpropertiesofthemembercross
section(e.g.,platebuckling)andtheEulerbucklingload(e.g.,flexuralbuckling).This
research project, sponsored by the American Iron and Steel Institute, extends the
appealinggeneralityofDSMtocoldformedsteelbeamsandcolumnswithperforations.

The elastic buckling properties of rectangular plates and coldformed steel beams

andcolumns,includingthepresenceofholes,arestudiedwiththinshellfiniteelement
eigenbuckling analysis. Buckled mode shapes unique to members with holes are
categorized. Parameter studies demonstrate that critical elastic buckling loads either
decreaseorincreasewiththepresenceofholes,dependingonthemembergeometryand
hole size, spacing, and location. Simplified alternatives to FE elastic buckling analysis
formemberswithholesaredevelopedwithclassicalplatestabilityequationsandfreely
availablefinitestripanalysissoftware.

ii

Experimentsoncoldformedsteelcolumnswithholesareconductedtoobservethe

interactionbetweenelasticbuckling,loaddeformationresponse,andultimatestrength.
The experimental results are used to validate an ABAQUS nonlinear finite element
protocol,whichisimplementedtosimulateloadingtocollapseofseveralhundredcold
formed steel beams and columns with holes. The results from these simulations,
supplemented with existing beam and column data, guide the development of design
equationsrelatingelasticbucklingandultimatestrengthforcoldformedsteelmembers
withholes.Theseequationsandthesimplifiedelasticbucklingpredictionmethodsare
presented as a proposed design procedure for an upcoming revision to the American
Iron and Steel Institutes North American Specification for the Design of ColdFormed
SteelStructuralMembers.

iii

SummaryofProgress
TheprimarygoalofthisAISIfundedresearchistoextendtheDirectStrengthMethodto
coldformedsteelmemberswithholes.

Researchbegins
September2005

ProgressReport#1
February2006

Accomplishments:
Evaluated the ABAQUS S9R5, S4, and S4R thin shell elements for accuracy and
versatilityinthinwalledmodelingproblems
Studiedtheinfluenceofelementaspectratioandelementquantitywhenmodeling
roundedcornersinABAQUS
DevelopedcustomMATLABtoolsformeshingholes,plates,andcoldformedsteel
membersinABAQUS
Determinedtheinfluenceofaslottedholeontheelasticbucklingofastructuralstud
channelandclassifiedlocal,distortional,andglobalbucklingmodes
Investigated the influence of hole size on the elastic buckling of a structural stud
channel
Performed a preliminary comparison of existing experimental data on coldformed
steelcolumnswithholestoDSMpredictions
Conducted a study on the influence of the hole width to plate width ratio on the
elasticbucklingbehaviorofasimplysupportedrectangularplate

Papersfromthisresearch:

Moen, C.D., Schafer, B.W. (2006) Impact of Holes on the Elastic Buckling of Cold
Formed Steel Columns with Application to the Direct Strength Method. Eighteenth
InternationalSpecialtyConferenceonColdFormedSteelStructures,Orlando,FL.

Moen,C.D.,Schafer,B.W.(2006)StabilityofColdFormedSteelColumnsWithHoles.
StabilityandDuctilityofSteelStructuresConference,Lisbon,Portugal.

ProgressReport#2
August2006

Accomplishments:
Evaluatedtheinfluenceofslottedholespacingontheelasticbucklingofplates(with
DeterminedtheimpactofflangeholesontheelasticbucklingofanSSMAstructural
stud

iv

Conducted a preliminary investigation into the nonlinear solution algorithms


availableinABAQUS
Compared the ultimate strength and loaddisplacement response of a rectangular
plate and an SSMA structural stud column with and without a slotted hole using
nonlinearfiniteelementmodelsinABAQUS
Calculatedtheeffectivewidthofarectangularplatewithandwithoutaslottedhole
usingnonlinearfiniteelementmodelsinABAQUS

ProgressReport#3
February2007

Accomplishments:
Conductedanexperimentalstudytoevaluatetheinfluenceofaslottedwebholeson
thecompressivestrength,ductility,andfailuremodesofshortandintermediatelength
Csectionchannelcolumns
Studiedtheinfluenceofslottedwebholesontheelasticbucklingbehaviorofcold
formed steel Csection beams and identified unique hole modes similar to those
observedincompressionmembers
Demonstrated that the Direct Strength Method is a viable predictor of ultimate
strengthforbeamswithholes

Papersfromthisresearch:

Moen, C.D., Schafer, B.W. (2008). Experiments on coldformed steel columns with
holes.ThinWalledStructures,46(10),11641182.

Moen,C.D.,Schafer,B.W.(2008).Observingandquantifyingtheelasticbucklingandt
estedresponseofcoldformedsteelcolumnswithholes.FifthInternationalConference
onThinWalledStructures,Brisbane,Australia.

ProgressReport#4
July2007

Accomplishments:

Completed experimental program including tensile coupon tests, elastic buckling


studyofspecimens,andDSMstrengthcomparison
Developed nonlinear finite element approach including a prediction method for
residual stresses and completed preliminary finite element studies of the 24 column
specimens
ProposedDSMapproachforcolumnswithholes,comparedtheoptionsagainstthe
columndatabase,conductedpreliminarysimulationstoexplorecolumnstrengthcurves

Papersfromthisresearch:
Moen,C.D.,Igusa,T.,Schafer,B.W.(2008).Predictionofresidualstressesandstrainsin
coldformedsteelmembers.ThinWalledStructures,46(11),12741289.

Moen,C.D.,Igusa,T.,Schafer,B.W.(2008).Amechanicsbasedpredictionmethodforr
esidualstressesandinitialplasticstrainsincoldformedsteelstructuralmembers.Fifth
ConferenceonCoupledInstabilitiesinMetalStructures,Sydney,Australia.

ProgressReport#5
February2008

Accomplishments:

Developed critical elastic buckling stress equations for stiffened and unstiffened
elementswithholesunderuniaxialcompression.
Implemented finite strip approximation method for predicting the critical elastic
localanddistortionalbucklingloadofcoldformedsteelmemberswithholes.
DerivedandtestedamethodforpredictingtheEulerbucklingloadsofcoldformed
steel columns and beams with holes, including flexural and flexuraltorsional
bucklingofcolumnsandlateraltorsionalbucklingofbeams.

Papers from this research:


Moen,C.D.,Schafer,B.W.(2008).Simplifiedmethodsforpredictingelasticbucklingof
coldformedsteelstructuralmemberswithholes.19thInternationalSpecialtyConference
onColdFormedSteelStructures,St.Louis,Missouri.

FinalReport

Fall2008

Accomplishments:

Created simulated column and beam experiment database with nonlinear finite
elementanalysisinABAQUSandCUFSMelasticbucklingapproximatemethods
Used the simulation results and existing column and beam experiment results to
develop and validate proposed DSM Holes design equations for CFS columns and
beams

vi

TableofContents

Chapter 1 Introduction............................................................................................................................. 1
Chapter 2 Thin-shell finite element modeling in ABAQUS ............................................................... 8
2.1 Comparison of ABAQUS thin-shell elements ............................................................................9
2.2 Modeling holes in ABAQUS ....................................................................................................... 16
2.3 Modeling Rounded Corners in ABAQUS ................................................................................. 18
2.4 Summary of modeling guidelines..............................................................................................21
Chapter 3 Elastic buckling of cold-formed steel cross-sectional elements with holes ..................22
3.1 Plate and hole dimensions .......................................................................................................... 23
3.2 Finite element modeling assumptions ......................................................................................25
3.3 Stiffened element in uniaxial compression ............................................................................... 25
3.4 Stiffened element in bending ...................................................................................................... 43
3.5 Unstiffened element in uniaxial compression ..........................................................................58
Chapter 4 Elastic buckling of cold-formed steel members with holes ............................................ 66
4.1 Finite element modeling assumptions ......................................................................................67
4.2 Elastic buckling of columns with holes ..................................................................................... 67
4.3 Elastic buckling of beams with holes....................................................................................... 126
Chapter 5 Experiments on cold-formed steel columns with holes ................................................ 159
5.1 Acknowledgements ................................................................................................................... 160
5.2 Testing Program ......................................................................................................................... 160
5.3 Elastic buckling calculations ..................................................................................................... 196
5.4 Experiment results ..................................................................................................................... 205
Chapter 6 Predicting residual stresses and plastic strains in cold-formed steel members .........223
6.1 Stress-strain coordinate system and notation ........................................................................ 226
6.2 Prediction method assumptions ..............................................................................................227
6.3 Derivation of the residual stress prediction method ............................................................. 230
6.4 Derivation of effective plastic strain prediction method ...................................................... 240
6.5 Employing the prediction method in practice: quantifying the coil radius influence ...... 244
6.6 Comparison of prediction method to measured residual stresses ...................................... 248
6.7 Discussion ................................................................................................................................... 254
6.8 Acknowledgements ................................................................................................................... 257
Chapter 7 Nonlinear finite element modeling of cold-formed steel structural members ...........258
7.1 Preliminary nonlinear FE studies ............................................................................................ 259
7.2 Nonlinear finite element modeling of columns with holes .................................................. 288
Chapter 8 The Direct Strength Method for cold-formed steel members with holes ................... 313
8.1 DSM for columns with holes .................................................................................................... 314
8.2 DSM for laterally braced beams with holes ............................................................................ 374
Chapter 9 Conclusions and proposed future work .......................................................................... 409
9.1 Conclusions ................................................................................................................................. 409
9.2 Future work ................................................................................................................................ 411
References .................................................................................................................................................. 415
Appendix A
ABAQUS input file generator in Matlab ................................................................ 419
Appendix B
ABAQUS element-based elastic buckling results .................................................. 425
Appendix C
Derivation of elastic buckling coefficients for unstiffened elements .................. 435
Appendix D
Elastic buckling prediction method of cross-sectional elements with holes......441
Appendix E
Derivation of global critical elastic buckling load for a column with holes.......445
Appendix F
Column experiment results ...................................................................................... 449
Appendix G
Residual stresses backstress for kinematic hardening implementation ........... 474
Appendix H
Experiment true stress-strain curves ....................................................................... 476
0H

42H

1H

43H

2H

4H

3H

45H

4H

46H

5H

47H

6H

48H

7H

49H

8H

450H

9H

451H

10H

452H

1H

453H

12H

45H

13H

45H

14H

456H

15H

457H

16H

458H

17H

459H

18H

460H

19H

461H

20H

462H

21H

463H

2H

46H

23H

465H

24H

46H

25H

467H

26H

468H

27H

469H

28H

470H

29H

471H

30H

472H

31H

473H

32H

47H

3H

475H

34H

476H

35H

47H

36H

478H

37H

479H

38H

480H

39H

481H

40H

482H

41H

483H

42H

48H

43H

485H

4H

486H

45H

487H

46H

48H

47H

489H

vii

Appendix
Appendix
Appendix
Appendix
Appendix
52H

51H

50H

49H

48H

I Column experiment nonlinear FE simulation results ............................................... 497


J Contact simulation in ABAQUS ................................................................................... 521
K
Simulated column experiments database ............................................................... 524
L
Simulated beam experiment database .................................................................... 528
M
Comparison of tested strengths to predicted strengths with AISI S100-07 ....... 531
490H

491H

492H

493H

49H

viii

ListofFigures

Figure 1.1 Perforations are provided in structural studs to accommodate utilities in the walls of
buildings................................................................................................................................... 2
Figure 1.2 Hole patterns in storage rack columns................................................................................... 2
Figure 1.3 Column elastic buckling curve generated with CUFSM ..................................................... 4
Figure 1.4 DSM global buckling failure design curve and equations .................................................. 4
Figure 1.5 DSM local buckling failure design curve and equations .....................................................5
Figure 1.6 DSM distortional buckling failure design curve and equations ......................................... 5
Figure 2.1 ABAQUS S4\S4R shell element with four nodes and a linear shape function, ABAQUS
S9R5 shell element with nine nodes and a quadratic shape function ............................10
Figure 2.2 Buckled shape of a stiffened plate ........................................................................................ 10
Figure 2.3 Accuracy of ABAQUS S9R5, S4, and S4R elements for a stiffened element with varying
aspect ratios, 8:1 finite element aspect ratio for the S9R5 element, 4:1 element aspect
ratio for the S4 and S4R elements .......................................................................................13
Figure 2.4 Accuracy of S4, S4R, and S9R5 elements as a function of the number of elements
provided per buckled half-wavelength, stiffened element, square waves (k=4).......... 14
Figure 2.5 Buckled shape of an unstiffened element, m=1 shown ...................................................... 15
Figure 2.6 Accuracy of S9R5 elements as the number of finite elements provided along an
unstiffened element varies, L/h=4 ...................................................................................... 15
Figure 2.7. Finite element mesh and plate dimensions: slotted, rectangular, and circular holes ..17
Figure 2.8 Hole discretization using S9R5 elements ............................................................................. 18
Figure 2.9 The critical elastic buckling stress converges to a constant magnitude when the S9R5
element aspect ratio a/b is between 0.5 and 2 and element corner angles are skewed
................................................................................................................................................. 18
Figure 2.10 ABAQUS S9R5 initial curvature limit requires at least five elements to model corner
................................................................................................................................................. 19
Figure 2.11 SSMA 600S162-68 C-section corner modeled with a) one S9R5 element, b) three S9R5
elements.................................................................................................................................. 20
Figure 2.12 The number of S9R5 corner elements has a minimal influence on the critical elastic
buckling loads of an SSMA 600162-68 C-section column with L=48 in. ........................ 21
Figure 3.1 Stiffened and unstiffened elements in a lipped C-section ................................................. 23
Figure 3.2 Element and hole dimension definitions ............................................................................. 24
Figure 3.3 Definition of unstiffened strip A and B for a plate with holes. .................................24
Figure 3.4 Definition of neutral axis location for stiffened elements in bending..............................24
Figure 3.5 ABAQUS boundary conditions and loading conditions for a stiffened element in
uniaxial compression ............................................................................................................ 26
Figure 3.6 Influence of a slotted hole on the elastic buckling stress of a simply supported
rectangular plate with varying length ............................................................................... 27
Figure 3.7 Comparison of buckled shape and displacement contours for a rectangular plate with
hhole/h=0.66 and L/Lhole =3, (a) with slotted hole and (b) without hole. Notice the
change in length and quantity of buckled cells with the addition of a slotted hole. ...28
Figure 3.8 Buckled shape of a simply supported plate (a) with a slotted hole and (b) without a
hole. L=15Lhole , hhole/h=0.66. The slotted hole dampens buckling but does not
significantly change the natural half-wavelength of the plate........................................ 28
Figure 3.9 (a) Slotted hole causes local buckling (hhole/h=0.26), compared to (b) buckled cells at
the natural half-wavelength of the plate............................................................................ 29
Figure 3.10 Definition of center-to-center dimension for the slotted holes ....................................... 30
53H

495H

54H

496H

5H

497H

56H

498H

57H

49H

58H

50H

59H

501H

60H

502H

61H

503H

62H

504H

63H

50H

64H

506H

65H

507H

6H

508H

67H

509H

68H

510H

69H

51H

70H

512H

71H

513H

72H

514H

73H

51H

74H

516H

75H

517H

76H

518H

7H

519H

78H

520H

79H

521H

80H

52H

ix

Figure 3.11 Influence of slotted hole spacing on the elastic buckling load of a long simply
supported rectangular plate ................................................................................................ 30
Figure 3.12 Comparison of buckled shapes for a long stiffened element (L=24 Lhole ) with a slotted
hole spacing of S/Lhole=4 and hhole/h=0.66, 0.44, and 0.26. ..................................................31
Figure 3.13 Variation in fcr with increasing hhole/h for a stiffened element correspond to buckling
mode shapes (see Figure 3.12 for examples of plate buckling and unstiffened strip
buckling mode shapes)......................................................................................................... 33
Figure 3.14 Unstiffened strip elastic buckling stress conversion from the net to the gross section
................................................................................................................................................. 35
Figure 3.15 Accuracy of stiffened element prediction method as a function of hole spacing S to
plate width h (a) without and (b) with the dimensional limits in Eq. (3.8) and Eq.(3.9)
................................................................................................................................................. 38
Figure 3.16 Accuracy of stiffened element prediction method as a function of hole spacing S to
length of hole Lhole (a) without and (b) with the dimensional limits in Eq. (3.8) and
Eq.(3.9) .................................................................................................................................... 38
Figure 3.17 Accuracy of the stiffened element prediction method as a function of hole width hhole
to plate width h (a) without and (b) with the dimensional limits in Eq. (3.8) and
Eq.(3.9) .................................................................................................................................... 38
Figure 3.18 For plates where the unstiffened strip is narrow compared to the plate width, plate
buckling occurs between the holes. ....................................................................................39
Figure 3.19 Plate buckling and unstiffened strip buckling may both exist for a plate with holes.
These modes are predicted conservatively as unstiffened strip buckling. ................... 40
Figure 3.20 Accuracy of prediction method for stiffened elements with square or circular holes as
a function of hole width hhole to plate width h. .................................................................. 40
Figure 3.21 Accuracy of the stiffened element elastic buckling prediction method as a function of
unstiffened strip width hstrip versus plate width h for offset holes (a) without and (b)
with the dimensional limits in Eq. (3.8) and Eq.(3.9) .......................................................41
Figure 3.22 Holes at the edge of a wide stiffened plate reduce the axial stiffness (and critical
elastic buckling stress) but do not change the buckled shape. ....................................... 42
Figure 3.23 Accuracy of the stiffened element elastic buckling prediction method as a function of
hole offset hole versus plate width h for offset holes (a) without and (b) with the
dimensional limits in Eq. (3.8), Eq.(3.9), and Eq. (3.11) ....................................................43
Figure 3.24 Boundary and loading conditions for a stiffened element in bending .......................... 43
Figure 3.25 Stiffened plates loaded with a linear bending stress gradient exhibit buckling of the
unstiffened strip adjacent to the hole in the compression region of the plate. .............45
Figure 3.26 Influence of slotted holes on critical elastic buckling stress fcr of stiffened elements in
bending as a function of (a) hole size relative to plate width and (b) hole spacing as a
function of hole length. ........................................................................................................ 46
Figure 3.27 Hole location influence on critical elastic buckling stress fcr for a stiffened plate in
bending (Y=0.50h) (Buckled mode shapes corresponding to A, B, C, and D are
provided in Figure 3.28.) ...................................................................................................... 47
Figure 3.28 The buckled mode shape changes as slotted holes move from the compression region
to the tension region of a stiffened element in bending (hhole/h=0.20). ...........................48
Figure 3.29 Hole location influence on critical elastic buckling stress fcr for a stiffened plate in
bending (Y=0.75h) ................................................................................................................. 49
Figure 3.30 Derivation of stress ratio for unstiffened strip A. ................................................... 51
Figure 3.31 Derivation B and conversion of the compressive stress at the edge of unstiffened
strip B to the stress fcrB at the edge of the plate ............................................................. 52
Figure 3.32 Derivation of fcrh for the case when hA+hholeY (when the hole is located partially in
the compressed region and partially in the tension region of the plate) ....................... 53
81H

523H

82H

524H

83H

52H

84H

526H

85H

527H

86H

528H

87H

529H

8H

530H

89H

531H

90H

532H

91H

53H

92H

534H

93H

53H

94H

536H

95H

537H

96H

538H

97H

539H

98H

540H

9H

541H

10H

542H

10H

543H

102H

54H

Figure 3.33 Derivation of fcrh for the case when hA+hhole<Y (hole lies completely in the compressed
region of the plate). ............................................................................................................... 54
Figure 3.34 Influence of Lhole/yA on the accuracy of the prediction method for stiffened elements in
bending (a) without and (b) with the dimensional limits defined in Eq. (3.9), Eq.
(3.24), Eq. (3.25), and Eq. (3.26)............................................................................................56
Figure 3.35 Influence of hA/Y on the accuracy of the prediction method for stiffened elements in
bending (a) without and (b) with the dimensional limits defined in Eq. (3.9), Eq.
(3.24), Eq. (3.25), and Eq. (3.26)............................................................................................56
Figure 3.36 Influence of S/h on the accuracy of the prediction method for stiffened elements in
bending (a) without and (b) with the dimensional limits defined in Eq. (3.9), Eq.
(3.24), Eq. (3.25), and Eq. (3.26)............................................................................................57
Figure 3.37 Influence of S/Lhole on the accuracy of the prediction method for stiffened elements in
bending (a) without and (b) with the dimensional limits defined in Eq. (3.9), Eq.
(3.24), Eq. (3.25), and Eq. (3.26)............................................................................................57
Figure 3.38 Influence of h/hhole on the accuracy of the prediction method for stiffened elements in
bending (a) without and (b) with the dimensional limits defined in Eq. (3.9), Eq.
(3.24), Eq. (3.25), and Eq. (3.26)............................................................................................58
Figure 3.39 ABAQUS boundary and loading conditions for unstiffened plate loaded uniaxially.
................................................................................................................................................. 58
Figure 3.40 The presence of holes causes a decrease in critical elastic buckling load for
unstiffened plates in uniaxial compression. ......................................................................60
Figure 3.41 Buckled shapes of unstiffened plates with holes. .............................................................60
Figure 3.42 The critical elastic buckling stress of a stiffened plate decreases as holes are shifted
toward the simply supported edge (+hole) ....................................................................... 62
Figure 3.43 The critical elastic buckling stress for stiffened elements with (a) transversely offset
holes and (b) centered holes (from Section 3.5.2) decreases as a function of hole
length Lhole to hA ..................................................................................................................... 62
Figure 3.44 (a) Comparison of ABAQUS and empirical plate buckling coefficients for an
unstiffened element with holes and (b) ABAQUS to predicted elastic buckling stress
for an unstiffened element ................................................................................................... 64
Figure 4.1 C-section and hole dimension notation ............................................................................... 68
Figure 4.2 Columns are modeled with pinned warping-free boundary conditions and
compressed from both ends ................................................................................................ 69
Figure 4.3(a) SSMA 250S162-33 web plate and structural stud, and (b) SSMA 400S162-33 web
plate and structural stud ...................................................................................................... 71
Figure 4.4. Effect of a slotted hole on the elastic buckling load of simply supported plates and
structural studs ...................................................................................................................... 72
Figure 4.5 The presence of a hole creates unique local buckling modes where unstiffened strip
buckling adjacent to the hole occurs symmetrically (LH) or asymmetrically (LH2)
increase the distortional tendency of the flanges.............................................................. 74
Figure 4.6 SSMA slotted hole location and local buckling LH mode, L=48 in.,
x/L=0.06,0.125,0.25,0.375,0.50. Note the distortional tendencies of the flanges at the
hole. ......................................................................................................................................... 74
Figure 4.7 Influence of a slotted hole on the (a) distortional (D) and (b) global flexural-torsional
(GFT) modes of a cold-formed steel column ..................................................................... 75
Figure 4.8 Influence of SSMA slotted hole location on Pcr for a 362S162-33 C-section (refer to
Figure 4.5, Figure 4.6, and Figure 4.7 for buckled shape summaries) ...........................76
Figure 4.9 Connection detail for structural stud to exterior wall requires a screw or bolt hole
placed in the stud flange (Western States Clay Products Association 2004) ................77
Figure 4.10 Influence of flange hole diameter on the local (L), distortional (D), and global (GFT)
elastic buckling loads of an SSMA 362S162-33 structural stud....................................... 78
103H

54H

104H

546H

105H

547H

106H

548H

107H

549H

108H

50H

109H

51H

10H

52H

1H

53H

12H

54H

13H

5H

14H

56H

15H

57H

16H

58H

17H

59H

18H

560H

19H

561H

120H

562H

12H

563H

12H

564H

123H

56H

124H

56H

xi

Figure 4.11 Local (L) buckling is dominated by flange and web deformation near the holes as
bhole/b exceeds 0.70 .................................................................................................................. 78
Figure 4.12 Experimental program boundary conditions as implemented in ABAQUS ................80
Figure 4.13 Influence of fixed-fixed boundary conditions versus warping free boundary
conditions on Pcrd for column experiments(L/H<4 ) as a function of (a) column length
to fundamental distortional half-wavelength calculated with CUFSM and (b) column
length to member length. .....................................................................................................87
Figure 4.14 Influence of fixed-fixed boundary conditions versus warping free boundary
conditions on Pcrl for column experiments ( L/H<4) as a function of (a) hole width
relative to column width and (b) hole length relative to column length ......................87
Figure 4.15 Influence of weak-axis pinned boundary conditions versus warping free boundary
conditions on (a) Pcrl as a function of hole length to column length and (b) Pcrd as a
function of column length to member length. .................................................................. 88
Figure 4.16 Rules for modeling a column net cross-section in CUFSM ............................................. 91
Figure 4.17 Local elastic buckling curve of net cross-section when (a) hole length is less than Lcrh
and (b) when hole length is greater than Lcrh..................................................................... 92
Figure 4.18 Comparison of CUFSM and ABAQUS predictions of unstiffened strip buckling. ......93
Figure 4.19 ABAQUS results verify CUFSM local buckling predictions for an SSMA 362S162-33
column with evenly spaced web holes. .............................................................................94
Figure 4.20 CUFSM and ABAQUS local buckling mode shapes are consistent when considering a
slotted flange hole. ................................................................................................................ 95
Figure 4.21 ABAQUS results verify CUFSM predictions for an SSMA 362S162-33 cross section
with evenly spaced flange holes. ........................................................................................95
Figure 4.22 ABAQUS predicts local plate buckling with distortional buckling interaction which is
not detected in CUFSM. ....................................................................................................... 96
Figure 4.23 ABAQUS results are slightly lower than CUFSM predictions, CUFSM predicts
correctly that plate local buckling controls over unstiffened strip buckling. ............... 97
Figure 4.24 Predicted Pcrh (CUFSM, buckling of the net cross-section) and Pcr (CUFSM, buckling
of the gross cross section, no hole) are compared relative to the ABAQUS Pcrl with
experiment boundary conditions as a function of (a) hole width to flat web width and
(b) hole length to column length ......................................................................................... 98
Figure 4.25 Predicted Pcrl (CUFSM approximate method) is compared relative to the ABAQUS
Pcrl with experiment boundary conditions as a function of (a) hole width to flat web
width and (b) hole length to column length...................................................................... 99
Figure 4.26 CUFSM approximate method for calculating Pcrd for a column with holes. ............... 101
Figure 4.27 Modified cross section to be used in CUFSM to predict Pcrd for a column with holes.
............................................................................................................................................... 102
Figure 4.28 ABAQUS boundary conditions and imposed rotations for web plate ........................ 103
Figure 4.29 Plate deformation from imposed edge rotations, hhole/h=0.50 ..................................... 104
Figure 4.30 Transverse rotational stiffness of the plate is significantly reduced in the vicinity of
the slotted hole .................................................................................................................... 104
Figure 4.31 Comparison of CUFSM and ABAQUS distortional buckling mode shapes. ..............107
Figure 4.32 CUFSM distortional buckling prediction method is conservative when considering
an SSMA 262S162-68 column with uniformly spaced holes. ........................................ 107
Figure 4.33 Warping-fixed boundary condition amplification of Pcrd .............................................. 109
Figure 4.34 Accuracy of the CUFSM approximate method for predicting Pcrd improves as column
length increases relative to the fundamental distortional half-wavelength for
warping-fixed columns ...................................................................................................... 110
Figure 4.35 A weighted thickness cross section can be input directly into a program that solves
the classical cubic stability equation for columns (e.g. CUTWP). ................................ 113
125H

567H

126H

568H

127H

569H

128H

570H

129H

571H

130H

572H

13H

573H

132H

574H

13H

57H

134H

576H

135H

57H

136H

578H

137H

579H

138H

580H

139H

581H

140H

582H

14H

583H

142H

584H

143H

58H

14H

586H

145H

587H

146H

58H

147H

589H

148H

590H

149H

591H

xii

Figure 4.36 Weak-axis flexural and flexural-torsional global buckling modes for an SSMA
1200S162-68 column with evenly spaced circular holes. ............................................... 114
Figure 4.37 Variation in net section properties as circular hole diameter increases. .....................115
Figure 4.38 Comparison of weighted thickness and weighted properties cross-sectional area.
............................................................................................................................................... 116
Figure 4.39 Comparison of weighted thickness and weighted properties strong axis moment
of inertia. .............................................................................................................................. 117
Figure 4.40 Comparison of weighted thickness and weighted properties weak axis moment
of inertia. .............................................................................................................................. 117
Figure 4.41 ABAQUS boundary conditions for warping free and applied unit twist at x=0 in. and
warping free but rotation restrained at x=100 in. ........................................................... 119
Figure 4.42 Angle of twist decreases linearly in the SSMA 1200S162-68 column with warping free
end conditions. .................................................................................................................... 119
Figure 4.43 The weighted properties approximation for Javg matches closely with the ABAQUS
prediction for the SSMA 12S00162-68 column with holes ............................................. 120
Figure 4.44 ABAQUS boundary conditions for warping free and applied unit twist at x=0 in. and
warping fixed and rotation restrained at x=100 in. ........................................................ 121
Figure 4.45 Angle of twist is nonlinear along the SSMA 1200S162-68 column with warping fixed
end conditions at x=100 in. ................................................................................................ 122
Figure 4.46 Comparison of weighted thickness and weighted properties approximations to
the ABAQUS derived warping torsion constant Cw,avg. ................................................ 123
Figure 4.47 Comparison of weighted thickness and weighted properties prediction methods
for the SSMA 1200S162-68 weak-axis flexural buckling mode. Predictions using net
section properties are also plotted as a conservative benchmark. ............................... 124
Figure 4.48 Comparison of weighted thickness and weighted properties prediction methods
for the SSMA 1200S162-68 flexural-torsional column buckling mode. Predictions
using net section properties are also plotted as a conservative benchmark. .............. 125
Figure 4.49 Cross section of beam specimen showing aluminum strap angles connected to Cflanges ................................................................................................................................... 127
Figure 4.50 C-section and hole dimension notation ........................................................................... 127
Figure 4.51 Experiment test setup with hole spacing, location of lateral bracing, spacing of
aluminum angle straps, and load points ......................................................................... 129
Figure 4.52 Finite element model boundary conditions for beam eigenbuckling analyses .......... 130
Figure 4.53 Channel and hole meshing details and modeling of aluminum angle straps ............ 131
Figure 4.54 ABAQUS meshing details for C-section rounded corners ............................................ 132
Figure 4.55 Modeling of the beam concentrated loads in ABAQUS ................................................ 133
Figure 4.56 Local buckling modes for specimen 2B,20,1&2(H) with and without holes .............. 135
Figure 4.57 Local buckling modes for specimen 3B,14,1&2(H) with and without holes .............. 135
Figure 4.58 Local buckling modes for specimen 6B,18,1&2(H) with and without holes .............136
Figure 4.59 Local buckling modes for specimen BP-40(H) with and without holes .....................136
Figure 4.60 Local buckling modes for specimen 12B,16,1&2(H) with and without holes ............ 137
Figure 4.61 Distortional buckling modes for specimen 2B,20,1&2(H) with and without holes ..138
Figure 4.62 Distortional buckling modes for specimen 3B,14,1&2(H) with and without holes ..139
Figure 4.63 Distortional buckling modes for specimen 6B,18,1&2(H) with and without holes ..139
Figure 4.64 Distortional buckling modes for specimen BP5-40(H) with and without holes ......... 140
Figure4.65Distortionalbucklingmodesforspecimen12B,16,1&2(H)withandwithoutholes 140
Figure 4.66 Elastic buckling curve for 12 deep specimen with modal participation summarized,
note that selected L and D are mixed local-distortional modes ................................... 141
Figure 4.67 Possible global buckling mode occurs about the compression flange lateral brace
point ...................................................................................................................................... 142
150H

592H

15H

593H

152H

594H

153H

59H

154H

596H

15H

597H

156H

598H

157H

59H

158H

60H

159H

601H

160H

602H

16H

603H

162H

604H

163H

605H

164H

60H

165H

607H

16H

608H

167H

609H

168H

610H

169H

61H

170H

612H

17H

613H

172H

614H

173H

615H

174H

61H

175H

617H

176H

618H

17H

619H

178H

620H

179H

621H

180H

62H

18H

623H

xiii

Figure 4.68 Influence of holes on beam specimen Mcrl (Channel 1 and Channel 2 plotted)
considering (a) all local buckling modes and (b) the lowest local buckling mode .... 147
Figure 4.69 Influence of holes on beam specimen Mcrd (Channel 1 and Channel 2 plotted) as a
function of hole depth to flat web depth considering (a) all distortional buckling
modes and (b) the lowest distortional buckling mode .................................................. 148
Figure 4.70 Influence of holes on beam specimen Mcrd (Channel 1 and Channel 2 plotted) as a
function of web depth to flange width considering (a) all distortional buckling modes
and (b) the lowest distortional buckling mode ............................................................... 148
Figure 4.71 Influence of test boundary conditions on Mcrl................................................................ 149
Figure 4.72 Influence of test boundary conditions on (a) Mcrd and (b) on the distortional halfwavelength........................................................................................................................... 151
Figure 4.73 Boost in Mcrd from the angle restraints increases as the fundamental distortional halfwavelength increases relative to the restraint spacing Sbrace .......................................... 151
Figure 4.74 Guidelines for restraining beam net cross-sections in the CUFSM local buckling
approximate method .......................................................................................................... 154
Figure 4.75 Comparison of ABAQUS to predicted Mcrl for C-sections with holes in the beam
database as a function of (a) web depth and (b) hole width relative to flat web depth
............................................................................................................................................... 155
Figure 4.76 Comparison of mechanics-based and weighted-average prediction methods to
ABAQUS results for the distortional buckling load Mcrd of C-sections with holes in
the elastic buckling database ............................................................................................. 156
Figure 4.77 ABAQUS boundary conditions and applied loading for an SSMA 1200S162-68 beam
with holes (hhole/h=0.50 shown) ........................................................................................ 157
Figure 4.78 Comparison of weighted thickness and weighted properties prediction methods
for the SSMA 1200S162-68 lateral-torsional beam buckling mode. Predictions using
net section properties are also plotted as a conservative benchmark. ......................... 158
Figure 5.1 Column testing parameters and naming convention....................................................... 161
Figure 5.2 Tested lengths of cold-formed steel columns with holes as a function of (a) column
length L and and (b) L versus out-to-out column width H ........................................... 163
Figure 5.3 Typical column specimens with slotted holes .................................................................. 164
Figure 5.4 Column test setup and instrumentation ............................................................................ 165
Figure 5.5 Novotechnik position transducer with ball-jointed magnetic tip .................................. 166
Figure 5.6 Central Machinery metal band saw used to rough cut column specimens .................. 167
Figure 5.7 362S162-33 short column specimen with bismuth end diaphragms .............................. 168
Figure 5.8 600S162-33 short column specimen oriented in CNC machine ...................................... 169
Figure 5.9 An end mill is used to prepare the column specimens .................................................... 169
Figure 5.10 The intermediate length specimens were end-milled in a manual milling machine. 170
Figure 5.11 The specimens are clamped at the webs only to avoid distortion of the cross-section
............................................................................................................................................... 170
Figure 5.12 Setup procedure for measuring specimen cross section dimensions .......................... 172
Figure 5.13 Procedure for measuring specimen cross-section dimensions ..................................... 173
Figure 5.14 Procedure for measuring flange-lip and flange-web angles ......................................... 174
Figure 5.15 Specimen measurement nomenclature ............................................................................ 175
Figure 5.16 Base metal and zinc thickness definitions ....................................................................... 177
Figure 5.17 Removal of tensile coupon zinc coating as a function of time ...................................... 179
Figure 5.18 A height gauge is used to measure specimen length ..................................................... 180
Figure 5.19 Lengths are measured at the four corners of the C-section column............................. 181
Figure 5.20 A dial gauge and precision stand are used to measure initial web imperfections ....184
Figure 5.21 Web imperfection measurement grid and coordinate system ...................................... 184
Figure 5.22 Column specimen alignment schematic .......................................................................... 186
182H

624H

183H

625H

184H

62H

185H

627H

186H

628H

187H

629H

18H

630H

189H

631H

190H

632H

19H

63H

192H

634H

193H

635H

194H

63H

195H

637H

196H

638H

197H

639H

198H

640H

19H

641H

20H

642H

201H

643H

20H

64H

203H

645H

204H

64H

205H

647H

206H

648H

207H

649H

208H

650H

209H

651H

210H

652H

21H

653H

21H

654H

213H

65H

214H

65H

xiv

Figure 5.23 Influence of platen bending stiffness on end moments for a fixed-fixed eccentric
column .................................................................................................................................. 188
Figure 5.24 Column specimen weak axis out-of-straightness schematic ......................................... 189
Figure 5.25 Digital calipers are used to measure the distance from the column web to platen edge
............................................................................................................................................... 190
Figure 5.26 Tensile coupons are first rough cut with a metal ban saw ............................................ 192
Figure 5.27 Tensile coupon dimensions as entered in the CNC milling machine computer ........ 192
Figure 5.28 A custom jig allows three tensile coupons to be milled at once in the CNC machine
............................................................................................................................................... 193
Figure 5.29 ATS machine used to test tensile coupons....................................................................... 194
Figure 5.30 Gradually yielding stress-strain curve with 0.2% strain offset method ...................... 195
Figure 5.31 Sharp-yielding stress strain curve using an autographic method for determining Fy
............................................................................................................................................... 195
Figure 5.32 (a) Local and distortional elastic buckled mode shapes for (a) short (L=48 in.) 362S16233 specimens and (b) intermediate length (L=48 in.) 362S162-33 specimens. ............. 198
Figure 5.33 Local and distortional elastic buckled mode shapes for (a) short (L=48 in.) 600S162-33
specimens and (b) intermediate length (L=48 in.) 600S162-33 specimens. .................. 198
Figure 5.34 Local (L) and distortional (D) DSM strength predictions are similar in magnitude for
both 362S162-33 and 600S162-33 cross-sections, indicating that L-D modal interaction
will occur during the tested response of the columns. .................................................. 201
Figure 5.35 Comparison of global mode shapes for intermediate length 362S162-33 and 600S16233 specimens. ....................................................................................................................... 205
Figure 5.36 Load-displacement progression for short column specimen 362-2-24-NH ................ 207
Figure 5.37 Load-displacement progression for short column specimen 362-2-24-H ..................... 208
Figure 5.38 Load-displacement curve for a 362S162-33 short column with, without a slotted hole
............................................................................................................................................... 208
Figure 5.39 Comparison of load-deformation response and lateral flange displacements for
specimen 362-2-24-NH ....................................................................................................... 210
Figure 5.40 Influence of a slotted hole on 362S162-33 short column lateral flange displacement 210
Figure 5.41 Load-displacement progression for short column specimen 600-1-24-NH ................ 212
Figure 5.42 Load-displacement progression for short column specimen 600-1-24-H .................... 212
Figure 5.43 Comparison of load-displacement response for short 600S162-33 column specimens
with and without holes ...................................................................................................... 213
Figure 5.44 Load-displacement progression, intermediate length column specimen 362-3-48-NH
............................................................................................................................................... 215
Figure 5.45 Load-displacement progression for intermediate length column specimen 362-3-48-H
............................................................................................................................................... 215
Figure 5.46 Load-displacement curve, 362S162-33 intermediate column with and without a hole
............................................................................................................................................... 216
Figure 5.47 362S162-33 long column mid-height flange displacements show the global torsional
failure mode ......................................................................................................................... 216
Figure 5.48 Load-displacement progression, intermediate length column specimen 600-1-48-NH
............................................................................................................................................... 218
Figure 5.49 Load-displacement progression, intermediate length column specimen 600-1-48-NH
............................................................................................................................................... 218
Figure 5.50 Load-displacement comparison of intermediate length 600S162-33 specimens with
and without holes ............................................................................................................... 219
Figure 5.51 Short 600S162-33 column flange-lip corner lifts off platen during post-peak portion of
test ......................................................................................................................................... 222
Figure 6.1 Cold-formed steel roll-forming: (left) Sheet coil enters roll-forming line, (right) steel
sheet is cold-formed into C-shape cross-section (photos courtesy of Bradbury Group).
............................................................................................................................................... 224
215H

657H

216H

658H

217H

659H

218H

60H

219H

61H

20H

62H

21H

63H

2H

64H

23H

65H

24H

6H

25H

67H

26H

68H

27H

69H

28H

670H

29H

671H

230H

672H

231H

673H

23H

674H

23H

675H

234H

67H

235H

67H

236H

678H

237H

679H

238H

680H

239H

681H

240H

682H

241H

683H

24H

684H

243H

685H

24H

68H

xv

Figure 6.2 Forming a bend: plastic bending and elastic springback of thin sheets results in a
nonlinear through-thickness residual stress distribution. ............................................. 225
Figure 6.3 Stress-strain coordinate system as related to the coiling and cold-forming processes.
............................................................................................................................................... 226
Figure 6.4 Roll-forming setup with sheet coil fed from the (a) top of the coil and (b) bottom of
coil. The orientation of the coil with reference to the roll-forming bed influences the
direction of the coiling residual stresses. ......................................................................... 229
Figure 6.5 Coiling of the steel sheet may result in residual curvature which results in bending
residual stresses as the sheet is flattened. ........................................................................ 231
Figure 6.6 Longitudinal residual stress distribution from coiling. ................................................... 232
Figure 6.7 Predicted longitudinal residual stress distribution from coiling, uncoiling, and
flattening of a steel sheet. ................................................................................................... 233
Figure 6.8 Cold-forming of a steel sheet. .............................................................................................. 235
Figure 6.9 Fully plastic transverse stress state from cold-forming. .................................................. 235
Figure 6.10 Force couple (Fpt) applied to simulate the elastic springback of the steel sheet after
the imposed radial deformation is removed. .................................................................. 236
Figure 6.11 Cold-forming of a steel sheet occurs as plastic bending and elastic springback,
resulting in a self-equilibrating transverse residual stress. ........................................... 237
Figure 6.12 Plastic bending and elastic springback from cold-forming in the transverse direction
result in longitudinal residual stresses because of the plane strain conditions.......... 238
Figure 6.13 Flowchart summarizing the prediction method for residual stresses in roll-formed
members. .............................................................................................................................. 239
Figure 6.14 Plastic strain distribution from sheet coiling with a radius less than elastic-plastic
threshold rep.......................................................................................................................... 241
Figure 6.15 Effective plastic strain in a cold-formed steel member from sheet coiling when the
radius rx is less than the elastic-plastic threshold rep....................................................... 242
Figure 6.16 Effective von Mises true plastic strain at the location of cold-forming of a steel sheet.
............................................................................................................................................... 243
Figure 6.17 Flowchart summarizing the prediction method for effective plastic strains in rollformed members ................................................................................................................. 244
Figure 6.18 Coil coordinate system and notation. ............................................................................... 245
Figure 6.19 Influence of sheet thickness and yield stress on through-thickness longitudinal
residual stresses (z-direction, solid lines are predictions for mean coil radius, dashed
lines for mean radius +/- one standard deviation). ....................................................... 247
Figure 6.20 The mean-squared error of the predicted and measured bending residual stresses for
de M. Batista and Rodrigues (De Batista and Rodrigues 1992), Specimen CP1 is
minimized when rx=1.60rinner. ............................................................................................ 251
Figure 6.21 (a) Histogram and (b) scattergram of bending residual stress prediction error (flat
cross-sectional elements) for 18 roll-formed specimens. ............................................... 254
Figure 6.22 (a) Histogram and (b) scattergram of bending residual stress prediction error (corner
cross-sectional elements) for 18 roll-formed specimens. ............................................... 254
Figure 6.23 Definition of apparent yield stress, effective residual stress, and effective plastic
strain as related to a uniaxial tensile coupon test. .......................................................... 256
Figure 7.1 True stress-strain curve derived from a tensile coupon test (Yu 2005)........................... 260
Figure 7.2 Simply supported boundary conditions with equation constraint coupling at loaded
edges ..................................................................................................................................... 261
Figure 7.3 Application of (a) uniform load and (b) uniform displacement to a stiffened element
............................................................................................................................................... 262
Figure 7.4 Type 1 imperfection (Schafer and Pekz 1998) ................................................................. 263
Figure 7.5 Modified Riks method load-displacement solutions and failure modes ...................... 265
Figure 7.6 Correlation between initial imperfection shape and fold line locations at failure ....... 265
245H

687H

246H

68H

247H

689H

248H

690H

249H

691H

250H

692H

251H

693H

25H

694H

253H

695H

254H

69H

25H

697H

256H

698H

257H

69H

258H

70H

259H

701H

260H

702H

261H

703H

26H

704H

263H

705H

264H

706H

265H

70H

26H

708H

267H

709H

268H

710H

269H

71H

270H

712H

271H

713H

27H

714H

xvi

Figure 7.7 Artificial damping load-displacement solutions and failure modes ............................. 268
Figure 7.8 Stiffened element boundary conditions with rigid body coupling at loaded edges ... 269
Figure 7.9 Initial geometric imperfection field used for the stiffened element with and without a
hole ........................................................................................................................................ 270
Figure 7.10 Deformation at ultimate load of a stiffened element with a hole loaded in
compression. The common failure mechanism is material yielding adjacent to the
hole followed by plate folding. ......................................................................................... 271
Figure 7.11 Load-displacement curve for the RIKS1 and RIKS2 models showing direction
reversal along load path ..................................................................................................... 272
Figure 7.12 RIKS1 and RIKS2 models experience convergence problems and return along the
loading path, the RIKS3 model successfully predicts a peak load and finds a postpeak load path ..................................................................................................................... 274
Figure 7.13 STATIC1 and STATIC2 load-displacement curves demonstrate convergence
difficulties near the peak load. .......................................................................................... 275
Figure 7.14 STAB1 and STAB2 load-displacement curves demonstrate a highly nonlinear postpeak equilibrium path ........................................................................................................ 277
Figure 7.15 The STAB1 and STAB2 models (artificial damping, displacement control) exhibit a
sharp drop in load as folding of the plate initiates near the hole. The STAB3 model
(artificial damping, load control) finds the compressive load at which a complete loss
of stiffness occurs. ............................................................................................................... 279
Figure 7.16 Comparison of ultimate limit state and elastic buckling plate behavior, initial
imperfections are not considered in these results .......................................................... 280
Figure 7.17 Load-displacement sensitivity to imperfection magnitude for a plate without a hole
............................................................................................................................................... 282
Figure 7.18 Load-displacement sensitivity to imperfection magnitude for a plate with a slotted
hole ........................................................................................................................................ 282
Figure 7.19 Calculation of effective width at a cross-section along a stiffened element ............ 284
Figure 7.20 Definition of longitudinal (S11) membrane stress .......................................................... 284
Figure 7.21 (a) longitudinal membrane stresses and (b) effective width of a stiffened element at
failure .................................................................................................................................... 285
Figure 7.22 (a) longitudinal membrane stresses and (b) effective width of a stiffened element with
a slotted hole at failure ....................................................................................................... 286
Figure 7.23 Effective width comparison for a plate with and without a slotted hole .................... 286
Figure 7.24 Through the thickness variation of effective width of a plate without a hole ............ 287
Figure 7.25 Through the thickness variation of effective width of a plate with a slotted hole .....287
Figure 7.26 Through thickness variation in longitudinal (S11) stresses in a plate at failure......... 288
Figure 7.27 ABAQUS boundary conditions simulating column experiments ................................ 290
Figure 7.28 ABAQUS plastic strain curve for specimen 362-1-24-NH assuming (a) plasticity
initiates at the proportional limit and (b) plasticity initiates at 0.2% offset yield stress
............................................................................................................................................... 293
Figure 7.29 ABAQUS plastic strain curve for specimen 600-1-24-NH assuming (a) plasticity
initiates at the proportional limit and (b) plasticity initiates at the beginning of the
yield plateau (refer to Appendix H for the details on the development of this curve).
............................................................................................................................................... 293
Figure 7.30 Influence of ABAQUS material model on the load-deformation response of specimen
600-1-24-NH (work this figure with Figure 7.29) ........................................................... 294
Figure 7.31 Slotted holes are filled with S9R5 elements to obtain no hole imperfection shapes ..295
Figure 7.32 L and D imperfection magnitudes described with a CDF (Schafer and Pekz 1998) 297
Figure 7.33 Method for measuring distortional imperfection magnitudes from experiments .....297
Figure 7.34 Definition of out-of-straightness imperfections implemented in ABAQUS ............... 298
Figure 7.35 ABAQUS element local coordinate system for use with residual stress definitions .300
Figure 7.36 Transverse residual stress distribution applied at the corners of the cross-section...300
273H

715H

274H

716H

275H

71H

276H

718H

27H

719H

278H

720H

279H

721H

280H

72H

281H

723H

28H

724H

283H

725H

284H

726H

285H

72H

286H

728H

287H

729H

28H

730H

289H

731H

290H

732H

291H

73H

29H

734H

293H

735H

294H

736H

295H

73H

296H

738H

297H

739H

298H

740H

29H

741H

30H

742H

301H

743H

302H

74H

xvii

Figure 7.37 Longitudinal residual stress distribution applied at the corners of the cross-section
............................................................................................................................................... 300
Figure 7.38 Equivalent plastic strain distribution at the corners of the cross-section .................... 301
Figure 7.39 Influence of section points on the unbalanced moment (accuracy) of the transverse
residual stress distribution as implemented in ABAQUS ............................................. 302
Figure 7.40 Load-displacement response of specimen 362-1-24-NH ............................................... 306
Figure 7.41 Load-displacement response of specimen 362-1-24-H ................................................... 306
Figure 7.42 Load-displacement response of specimen 362-1-48-NH ............................................... 307
Figure 7.43 Load-displacement response of specimen 362-1-48-H ................................................... 307
Figure 7.44 Load-displacement response of specimen 600-1-24-NH ............................................... 308
Figure 7.45 Load-displacement response of specimen 600-2-24-H ................................................... 308
Figure 7.46 Load-displacement response of specimen 600-1-48-NH ............................................... 309
Figure 7.47 Load-displacement response of specimen 600-3-48-H ................................................... 309
Figure 7.48 Influence of residual stresses (RS) and plastic strains (PS) on the FE loaddisplacement response of specimen 600-1-24-NH .......................................................... 311
Figure 7.49 Influence of residual stresses (RS) and plastic strains (PS) on the FE loaddisplacement response of specimen 362-1-24-NH. ......................................................... 312
Figure 8.1 ABAQUS simulated column experiments boundary conditions and application of
loading .................................................................................................................................. 315
Figure 8.2 SSMA 800S250-97 structural stud with web holes considered in the DSM distortional
buckling study ..................................................................................................................... 318
Figure 8.3 SSMA 800S250-97 structural stud failure mode transition from distortional buckling to
yielding at the net section .................................................................................................. 320
Figure 8.4 Comparison of simulated column strengths (Anet/Ag=1.0) to (a) the existing DSM
distortional buckling design curve and to (b) the proposed DSM distortional
buckling curve for columns with holes ............................................................................ 321
Figure 8.5 Comparison of simulated column strengths (Anet/Ag=0.90) to (a) the existing DSM
distortional buckling design curve and to (b) the proposed DSM distortional
buckling curve for columns with holes ............................................................................ 321
Figure 8.6 Comparison of simulated column strengths (Anet/Ag=0.80) to (a) the existing DSM
distortional buckling design curve and to (b) the proposed DSM distortional
buckling curve for columns with holes ............................................................................ 322
Figure 8.7 Comparison of simulated column strengths (Anet/Ag=0.70) to (a) the existing DSM
distortional buckling design curve and to (b) the proposed DSM distortional
buckling curve for columns with holes ............................................................................ 322
Figure 8.8 Comparison of simulated column strengths (Anet/Ag=0.60) to (a) the existing DSM
distortional buckling design curve and to (b) the proposed DSM distortional
buckling curve for columns with holes ............................................................................ 322
Figure 8.9 Comparison of simulated column strengths (Anet/Ag=1.00) to (a) the existing DSM
global buckling design curve and to (b) the proposed DSM distortional buckling
curve for columns with holes ............................................................................................ 325
Figure 8.10 Comparison of simulated column strengths (Anet/Ag=0.90) to (a) the existing DSM
global buckling design curve and to (b) the proposed DSM distortional buckling
curve for columns with holes ............................................................................................ 326
Figure 8.11 Comparison of simulated column strengths (Anet/Ag=0.80) to (a) the existing DSM
global buckling design curve and to (b) the proposed DSM distortional buckling
curve for columns with holes ............................................................................................ 326
Figure 8.12 Comparison of column test-to-prediction ratios for columns (Anet/Ag=1.0) failing by
local-global buckling interaction as a function of (a) local slenderness (b) global
slenderness ........................................................................................................................... 328
30H

745H

304H

746H

305H

74H

306H

748H

307H

749H

308H

750H

309H

751H

310H

752H

31H

753H

312H

754H

31H

75H

314H

756H

315H

75H

316H

758H

317H

759H

318H

760H

319H

761H

320H

762H

321H

763H

32H

764H

32H

765H

324H

76H

325H

76H

326H

768H

327H

769H

xviii

Figure 8.13 Comparison of column test-to-prediction ratios for columns failing by local-global
buckling interaction with Pne calculated (a) without the influence of holes (b) and
with the influence of holes ................................................................................................. 330
Figure 8.14 Comparison of column test-to-prediction ratios for columns failing by local-global
buckling interaction as a function of Pynet/Pne where Pne is calculated (a) without the
influence of holes (b) and with the influence of holes ................................................... 331
Figure 8.15 SSMA 350S162-68 column failure mode changes from distortional-flexural torsional
buckling failure to weak axis flexure as hole size increases (L=34 in.) ........................ 331
Figure 8.16 SSMA 800S250-43 (L=74 in.) column web local buckling changes to unstiffened strip
buckling at peak load as hole size increases.................................................................... 332
Figure 8.17 Comparison of DSM local buckling design curve options when Pynet=0.80 Pyg and (a)
Pcre=100Pyg, (b) Pcre=5Pyg, and (c) Pcre=Pyg .......................................................................... 334
Figure 8.18 Comparison of simulated test strengths to predictions for columns with local
buckling-controlled failures as a function of local slenderness (tested strength is
normalized by Pne)............................................................................................................... 345
Figure 8.19 Comparison of simulated test strengths to predictions for columns with local
buckling-controlled failures as a function of local slenderness (tested strength is
normalized by Pyg) .............................................................................................................. 346
Figure 8.20 Comparison of simulated test strengths to predictions for columns with distortional
buckling-controlled failures as a function of distortional slenderness ........................347
Figure 8.21 Comparison of simulated test strengths to predictions for columns with global
buckling-controlled failures (i.e., no local interaction) as a function of global
slenderness ........................................................................................................................... 348
Figure 8.22 Test-to-predicted ratios for local buckling-controlled simulated column failures as a
function of local slenderness ............................................................................................. 349
Figure 8.23 Test-to-predicted ratios for distortional buckling-controlled simulated column
failures as a function of distortional slenderness ........................................................... 350
Figure 8.24 Test-to-predicted ratios for simulated global buckling-controlled column failures (i.e.,
no local buckling interaction) as a function of global slenderness ............................... 351
Figure 8.25 Test-to-predicted ratios for simulated local buckling-controlled column failures as a
function of net cross-sectional area to gross cross-sectional area ................................. 352
Figure 8.26 Test-to-predicted ratios for simulated distortional buckling-controlled column
failures as a function of net cross-sectional area to gross cross-sectional area ...........353
Figure 8.27 Test-to-predicted ratios for simulated global buckling-controlled column failures as a
function of net cross-sectional area to gross cross-sectional area ................................. 354
Figure 8.28 Test-to-predicted ratios for simulated local buckling-controlled column failures as a
function of column length, L, to flat web width, h ......................................................... 355
Figure 8.29 Test-to-predicted ratios for simulated distortional buckling-controlled column
failures as a function of column length, L, to flat web width, h.................................... 356
Figure 8.30 Test-to-predicted ratios for simulated global buckling-controlled column failures as a
function of column length, L, to web width, h ................................................................ 357
Figure 8.31 Comparison of experimental test strengths to predictions for columns with local
buckling-controlled failures as a function of local slenderness (tested strength is
normalized by Pne)............................................................................................................... 360
Figure 8.32 Comparison of experimental test strengths to predictions for columns with local
buckling-controlled failures as a function of local slenderness (tested strength is
normalized by Py) ................................................................................................................ 361
Figure 8.33 Comparison of experimental test strengths to predictions for columns with
distortional buckling-controlled failures as a function of distortional slenderness .. 362
Figure 8.34 Comparison of experimental test strengths to predictions for columns with global
buckling-controlled failures as a function of global slenderness ................................. 363
328H

70H

329H

71H

30H

72H

31H

73H

32H

74H

3H

75H

34H

76H

35H

7H

36H

78H

37H

79H

38H

780H

39H

781H

340H

782H

341H

783H

342H

784H

34H

785H

34H

786H

345H

78H

346H

78H

347H

789H

348H

790H

349H

791H

xix

Figure 8.35 Test-to-predicted ratios for experiment local buckling-controlled column failures as a
function of local slenderness ............................................................................................. 364
Figure 8.36 Test-to-predicted ratios for experiment distortional buckling-controlled column
failures as a function of distortional slenderness .......................................................... 365
Figure 8.37 Test-to-predicted ratios for experiment global buckling-controlled column failures as
a function of global slenderness ........................................................................................ 366
Figure 8.38 Test-to-predicted ratios for experiment local buckling-controlled column failures as a
function of net cross-sectional area to gross cross-sectional area ................................. 367
Figure 8.39 Test-to-predicted ratios for experiment distortional buckling-controlled column
failures as a function of net cross-sectional area to gross cross-sectional area ...........368
Figure 8.40 Test-to-predicted ratios for experiment global buckling-controlled column failures as
a function of net cross-sectional area to gross cross-sectional area .............................. 369
Figure 8.41 Test-to-predicted ratios for experiment local buckling-controlled column failures as a
function of column length, L, to flat web width, h ......................................................... 370
Figure 8.42 Test-to-predicted ratios for experiment distortional buckling-controlled column
failures as a function of column length, L, to flat web width, h.................................... 371
Figure 8.43 Test-to-predicted ratios for experiment global buckling-controlled column failures as
a function of column length, L, to web width, h ............................................................. 372
Figure 8.44 ABAQUS simulated beam experiments boundary conditions and application of
loading .................................................................................................................................. 375
Figure 8.45 SSMA 800S162-43 beam with web holes considered in the DSM local buckling study
............................................................................................................................................... 378
Figure 8.46 Comparison of simulated beam strengths (Inet/Ig=1.0, no holes) to (a) the existing DSM
local buckling design curve and to (b) the proposed DSM local buckling curve for
beams with holes ................................................................................................................. 380
Figure 8.47 Comparison of simulated beam strengths (Inet/Ig=0.95) to (a) the existing DSM local
buckling design curve and to (b) the proposed DSM local buckling curve for beams
with holes ............................................................................................................................. 380
Figure 8.48 Comparison of simulated beam strengths (Inet/Ig=0.90) to (a) the existing DSM local
buckling design curve and to (b) the proposed DSM local buckling curve for beams
with holes ............................................................................................................................. 381
Figure 8.49 Comparison of simulated beam strengths (Inet/Ig=0.85) to (a) the existing DSM local
buckling design curve and to (b) the proposed DSM local buckling curve for beams
with holes ............................................................................................................................. 381
Figure 8.50 SSMA 550S162-54 structural stud failure mode transition from distortional buckling
to yielding at the net section .............................................................................................. 383
Figure 8.51 Comparison of simulated beam strengths (Inet/Ig=1.0) to (a) the existing DSM
distortional buckling design curve and to (b) the proposed DSM distortional
buckling curve for beams with holes ............................................................................... 384
Figure 8.52 Comparison of simulated beam strengths (Inet/Ig=0.95) to (a) the existing DSM
distortional buckling design curve and to (b) the proposed DSM distortional
buckling curve for beams with holes ............................................................................... 384
Figure 8.53 Comparison of simulated beam strengths (Inet/Ig=0.90) to (a) the existing DSM
distortional buckling design curve and to (b) the proposed DSM distortional
buckling curve for beams with holes ............................................................................... 385
Figure 8.54 Comparison of simulated test strengths to predictions for laterally braced beams with
local buckling-controlled failures as a function of local slenderness ........................... 394
Figure 8.55 Comparison of simulated test strengths to predictions for laterally braced beams with
distortional buckling-controlled failures as a function of distortional slenderness .. 395
Figure 8.56 Test-to-predicted ratios for local buckling-controlled simulated laterally braced beam
failures as a function of local slenderness........................................................................ 396
350H

792H

351H

793H

352H

794H

35H

795H

354H

796H

35H

79H

356H

798H

357H

79H

358H

80H

359H

801H

360H

802H

361H

803H

362H

804H

36H

805H

364H

806H

365H

807H

36H

80H

367H

809H

368H

810H

369H

81H

370H

812H

371H

813H

xx

Figure 8.57 Test-to-predicted ratios for distortional buckling-controlled simulated laterally


braced beam failures as a function of distortional slenderness .................................... 397
Figure 8.58 Test-to-predicted ratios for simulated local buckling-controlled laterally braced beam
failures as a function of net cross-sectional moment of inertia to gross cross-sectional
moment of inertia ................................................................................................................ 398
Figure 8.59 Test-to-predicted ratios for simulated distortional buckling-controlled laterally
braced beam failures as a function of net cross-sectional moment of inertia to gross
cross-sectional moment of inertia ..................................................................................... 399
Figure 8.60 Test-to-predicted ratios for simulated local buckling-controlled laterally braced beam
failures as a function of column length, L, to beam depth, H ....................................... 400
Figure 8.61 Test-to-predicted ratios for simulated distortional buckling-controlled laterally
braced beam failures as a function of column length, L, to H ....................................... 401
Figure 8.62 Comparison of experimental test strengths to predictions for laterally braced beams
with local buckling-controlled failures as a function of local slenderness .................. 404
Figure 8.63 Comparison of experimental test strengths to predictions for laterally braced beams
with distortional buckling-controlled failures as a function of distortional slenderness
............................................................................................................................................... 405
Figure 8.64 Test-to-predicted ratios for experimental local buckling-controlled laterally braced
beam failures as a function of net cross-sectional moment of inertia to gross crosssectional moment of inertia ............................................................................................... 406
Figure 8.65 Test-to-predicted ratios for experimental distortional buckling-controlled laterally
braced beam failures as a function of net cross-sectional moment of inertia to gross
cross-sectional moment of inertia ..................................................................................... 407
Figure C.1 CUFSM finite strip modeling definition for an unstiffened element in compression 435
Figure C.2 The plate buckling coefficient kA for an unstiffened element in compression (the
multiple curves represent 0A1 with a step of 0.1, 11 curves total) ......................... 436
Figure C.3 The fitted curve for kA is a conservative predictor when Lhole/yA2 ............................... 437
Figure C.4 CUFSM finite strip modeling definition for an unstiffened element with compression
on the free edge, tension on the simply-supported edge .............................................. 438
Figure C.5 Variation in plate buckling coefficient kB for an unstiffened element with B ranging
from 0 to 10 .......................................................................................................................... 438
Figure C.6 Curve fit to minimum kB for B ranging from 0 to 10 ..................................................... 439
Figure C.7 Family of curves used to simulate boost in kB when Lhole/hB2, B ranges from 0 to 10
............................................................................................................................................... 440
Figure E.1 Long column with two holes spaced symmetrically about the longitudinal midline.445
372H

814H

37H

815H

374H

816H

375H

817H

376H

81H

37H

819H

378H

820H

379H

821H

380H

82H

381H

823H

382H

824H

38H

825H

384H

826H

385H

827H

386H

82H

387H

829H

38H

830H

xxi

ListofTables

Table 3.1 Plate widths corresponding to SSMA structural stud designations ..................................................... 26
Table 3.2 Parameter ranges in stiffened element verification study. .................................................................... 36
Table 3.3 Parameter range for stiffened element verification study with offset holes. ...................................... 41
Table 3.4 Parameter ranges considered for stiffened elements in bending with holes. ...................................... 44
Table 3.5 Study parameter limits for stiffened element in bending (Y/h=0.50) with offset holes .................... 46
Table 3.6 Parameter range for study of regularly-spaced holes on unstiffened elements. ................................ 59
Table 3.7 Parameter range considered for unstiffened element study with offset holes ................................... 61
Table 4.1 SSMA structural stud and plate dimensions ........................................................................................... 70
Table 4.2 Summary of column experimental data .................................................................................................. 79
Table 4.3 Fixed-fixed column experiment dimensions and material properties................................................. 82
Table 4.4 Fixed-fixed column experiment elastic buckling properties ................................................................. 83
Table 4.5 Weak-axis pinned column experiment dimensions and material properties ..................................... 84
Table 4.6 Weak-axis pinned column experiment elastic buckling properties ..................................................... 84
Table 4.7 Parameter ranges for fixed-fixed and weak-axis pinned column specimens with holes .................. 85
Table 4.8 DSM prequalification limits for C-sections ............................................................................................. 85
Table 4.9 DSM prequalification limits for beam C-sections................................................................................. 143
Table 4.10 Parameter ranges for beam specimens with holes ............................................................................. 143
Table 4.11 Beam experiment cross-section dimensions, material properties, and tested strengths ............... 144
Table 4.12 Beam experiment elastic buckling results ........................................................................................... 145
Table 5.1 FSM local and distortional buckling half-wavelengths for nominal 362S162-33 and 600S162-33
cross-sections.................................................................................................................................................... 162
Table 5.2 Voltage conversion factors for column test instrumentation .............................................................. 166
Table 5.3 Summary of measured cross section dimensions................................................................................. 175
Table 5.4 Summary of measured lip-flange and flange-web cross section angles ............................................ 176
Table 5.5 Specimen bare steel and zinc coating thicknesses ................................................................................ 178
Table 5.6 Measured column specimen length ....................................................................................................... 181
Table 5.7 Specimen end flatness .............................................................................................................................. 182
Table 5.8 Measured slotted hole dimensions and locations ................................................................................ 183
Table 5.9 Initial web imperfections (deviations from the average elevation of the web) ................................ 185
Table 5.10 Specimen gross centroid and offset from applied load during tests ............................................... 187
Table 5.11 Summary of out-of-straightness calculations ..................................................................................... 190
Table 5.12 Voltage conversion factors for tensile coupon testing ....................................................................... 194
Table 5.13 Summary of column specimen steel yield stress ................................................................................ 195
Table 5.14 Column specimen steel yield stress statistics...................................................................................... 196
Table 5.15 Critical elastic buckling loads, influence of holes on elastic buckling ............................................. 200
Table 5.16 Specimen ultimate strength results ...................................................................................................... 206
Table 6.1 Statistics of the residual stresses in roll-formed members .................................................................. 249
Table 6.2 Radial location in the coil that minimizes the sum of the mean square prediction error for rollformed members .............................................................................................................................................. 252
Table 7.1 Summary of nonlinear finite element models and associated solution controls .............................. 271
Table 7.2 Local and distortional imperfection magnitudes ................................................................................. 297
Table 7.3 Out-of-straightness imperfection magnitudes ...................................................................................... 298
Table 7.4 Comparison of nonlinear FE simulation peak loads to experiments ................................................. 303
Table 8.1 DSM test-to-predicted statistics for column simulations .................................................................... 344
Table 8.2 DSM test-to-predicted ratio statistics for column experiments .......................................................... 359
Table 8.3 DSM test-to-predicted ratio statistics for column experiments (stub columns only) ....................... 359
Table 8.4 DSM test-to-predicted statistics for laterally braced beam simulations ............................................ 393
Table 8.5 DSM test-to-predicted ratio statistics for column experiments .......................................................... 403
389H

831H

390H

832H

391H

83H

392H

834H

39H

835H

394H

836H

395H

837H

396H

83H

397H

839H

398H

840H

39H

841H

40H

842H

401H

843H

402H

84H

403H

845H

40H

846H

405H

847H

406H

84H

407H

849H

408H

850H

409H

851H

410H

852H

41H

853H

412H

854H

413H

85H

41H

856H

415H

857H

416H

85H

417H

859H

418H

860H

419H

861H

420H

862H

421H

863H

42H

864H

423H

865H

42H

86H

425H

867H

426H

86H

427H

869H

428H

870H

429H

871H

430H

872H

431H

873H

432H

874H

43H

875H

43H

876H

xxii

Chapter 1

Introduction
0B

The goal of this research work is to develop a general design method for cold

formedsteelstructuralmemberswithholes.Coldformedsteelbeamsandcolumnsare
typically manufactured with perforations. For example, in low and midrise
construction, holes are prepunched in structural studs to accommodate the passage of
utilities in the walls and ceilings of buildings as shown in Figure 1.1. In coldformed
87H

steelstoragerackcolumns,perforationpatternsareprovidedtoallowforvariableshelf
configurations as shown in Figure 1.2. (Members with discrete holes, for example C
87H

sectionswithpunchedholesasshownin Figure1.1,aretheresearchfocusinthisthesis,
879H

althoughmanyofthetoolsandmethodsdevelopedherecanbeextendedtoperforation
patternsinstoragerackswithadditionalresearcheffort.)Existingdesignproceduresfor
coldformed steel members with holes are limited to certain hole sizes, shapes, and
configurations. These limitations can hamper an engineers design flexibility and
decrease the reliability of coldformed steel components where holes exceed these
prescriptivelimits.


Figure1.1Perforationsareprovidedinstructuralstudstoaccommodateutilitiesinthewallsofbuildings

Figure1.2Holepatternsinstoragerackcolumns

ThebasicframeworkofthedesignproceduredevelopedinthisthesisistheDirect

Strength Method (DSM) (AISIS100 2007, Appendix 1). DSM is relatively new and
represents a major advancement in coldformed steel design because it provides
engineersandcoldformedsteelmanufacturerswiththetoolstopredictthestrengthofa
memberwithanygeneralcrosssection.Coldformedsteelmembersaremanufactured
from thin sheet steel, and therefore member resistance is influenced by crosssection
instabilities(e.g.,platebucklinganddistortionofopencrosssections)inadditiontothe
globalbucklinginfluenceconsideredinthickerhotrolledsteelsections.DSMexplicitly
defines the relationship between elastic buckling and loaddeformation response with
empiricalequationstopredictultimatestrength.

To calculate the capacity of a coldformed steel member with DSM, the elastic

buckling properties of a general coldformed steel crosssection are obtained from an


elasticbucklingcurve.Thecurvecanbegeneratedwithsoftwareemployingthefinite
stripmethodtoperformaseriesofeigenbucklinganalysesoverarangeofbuckledhalf
wavelengths. In this research work the freely available program CUFSM is utilized
(Schafer and dny 2006). An example of an elastic buckling curve is provided in
Figure1.3foracoldformedsteelCsectioncolumnandhighlightsthethreecategoriesof
80H

elastic buckling considered in DSM local buckling, distortional buckling, and global
buckling. Local buckling occurs as plate buckling of individual slender elements in a
crosssection. Distortional buckling exists only for open crosssections such as a C
section,wherethecompressedflangesbuckleinwardoroutwardalongthelengthofa
member. Global buckling, also known as Euler buckling, defines buckling of the full
member at long halfwavelengths including both flexural and flexuraltorsional effects
(andlateraltorsionaleffectsinbeams).

The critical elastic buckling loads associated with local, distortional, and global

bucklingPcrl,Pcrd,andPcreforcolumns(Mcrl,Mcrd,andMcreforbeams),canbeobtained
directlyfromtheelasticbucklingcurve.Thecriticalelasticbucklingloadsarethenused
topredicttheultimatestrengthwiththreeempiricaldesigncurvespresentedin Figure
81H

1.4 to Figure 1.6 for coldformed steel columns. (The current DSM column design
82H

equations for members without holes are also provided in these figures.) The local,
distortional, and global slenderness of a member (l, d, c) are calculated from the
criticalelasticbucklingloads,definingamemberssensitivitytoeachtypeofbucklingat

failure (high slenderness corresponds to high sensitivity, low slenderness to low


sensitivity). The nominal resistances (Pnl, Pnd, and Pne) are obtained by inserting the
slenderness magnitudes into the DSM design equations. The minimum of the local,
distortional,andglobalnominalstrengthsistakenasthestrengthofthemember.

20
18

P (kips)

loadcrfactor

Distortional
buckling

Local buckling

16
14
12
10

PPcrd

Global
buckling

crd

6
4

Pcre

Pcrl P

crl

2
0

10

Pcre

10

10

10

half-wavelength

half-wavelength (in.)

Figure1.3ColumnelasticbucklingcurvegeneratedwithCUFSM

1.5
Flexural, Torsional, or Torsional-Flexural Buckling
The nominal axial strength, Pne, for flexural or torsional- flexural buckling is
2
for c 1.5 Pne = 0.658 c Py

1
Pne /Py

0.877
for c > 1.5 Pne = 2 Py = 0.877 Pcre

c

where

Py

= AgFy

Py Pcre

Pcre= Critical elastic global column buckling load


Ag

0.5

0.5

= gross area of the column

1.5

2.5

3.5

0.5

Global slenderness, c =(Py /P cre)

Figure1.4DSMglobalbucklingfailuredesigncurveandequations

1.5
Local Buckling
The nominal axial strength, Pnl, for local buckling is
for l 0.776 Pnl = Pne

P
for l > 0.776 Pnl = 1 0.15 crl

Pne

where

0.4

P
crl
Pne

0.4

Pne

Pne Pcrl

Pn /Pne
l

Pcrl = Critical elastic local column buckling load

Pne =

Nominal axial strength for global buckling

0.5

Global
failure
0

0.5

Local buckling interacts with


global buckling at failure

1.5

2.5

3.5

local slenderness, =(Pne/P cr )0.5


l
l

Figure1.5DSMlocalbucklingfailuredesigncurveandequations

1.5
Distortional Buckling
The nominal axial strength, Pnd, for distortional buckling is
for d 0.561 Pnd = Py

for d > 0.561 Pnd = 1 0.25 crd


Py

1
Pnd /Py

where

0.6

Pcrd
P
y

0.6

Py

= Py Pcrd

Pcrd = Critical elastic distortional column buckling load


Py = Column yield strength

0.5

0.5

1.5

2.5

3.5

Distortional slenderness, d=(Py /Pcrd)0.5

Figure1.6DSMdistortionalbucklingfailuredesigncurveandequations

This research aims to extend the appealing generality of DSM to coldformed steel

memberswithperforations.Theprimaryresearchgoalsaretostudyandquantifythe
influenceofholesontheelasticbucklingofcoldformedsteelbeamsandcolumnsand
thentodevelopmodificationstotheexistingDSMdesignequationswhichrelateelastic
bucklingtoultimatestrength.Theresearchplanisimplementedinthreephases:
PhaseI(Chapters24)
1. Studytheinfluenceofholesontheelasticbucklingofthinplates,andthenoncold
formedsteelbeamsandcolumns.

2. Evaluate the viability of DSM for members with holes by comparing existing
experimentsonmemberswithholestothecurrentDSMspecification.

PhaseII(Chapters57)
1.Conductexperimentsoncoldformedsteelcolumnstoobservetheinfluenceofholes
onultimatestrengthandpostbucklingresponse.

2.Defineandvalidateanonlinearfiniteelementmodelingprotocolthroughparameter
studiesonthinplatesandcomparisontoexperimentalresults.

PhaseIII(Chapters78)
1. Formalize the relationship between elastic buckling and ultimate strength for
memberswithholesusingnonlinearfiniteelementsimulationsandexistingdata.

2.ModifythecurrentDSMspecificationtoaccountformemberswithholes

Phase I research is primarily focused on elastic buckling. Chapter 2 describes

preliminary thin shell finite element eigenbuckling studies which are used to evaluate
the accuracy of different shell element types in ABAQUS and to define finite element
meshingguidelines.Chapter3extendsthiselasticbucklingresearchwitheigenbuckling
analysesoftypicalcrosssectionalelementsconsideredincoldformedsteeldesign.For

example, a stiffened element is a simplysupported plate used to model the web of a


coldformedsteelCsectionandanunstiffenedelementisaplatesimplysupportedon
threeedgesandfreeonthefourthedgetosimulatethebehaviorofthefreelegofacold
formed steel hat section. Chapter 4 examines the elastic buckling of full coldformed
steelbeamsandcolumnswithholesanddevelopsusefulsimplifiedmethodstopredict
elastic buckling, including the influence of holes, without finite element analysis. The
elasticbucklingpropertiesofexistingbeamandcolumnexperimentsarealsocalculated
andmergedwiththetestedstrengthsintoadatabase.Thisdatabaseisemployednear
the end of the project to validate the proposed modifications to the DSM design
equationsformemberswithholes.

PhaseIImarksashiftfromelasticbucklingtothestudyoftheinfluenceofholeson

loaddeformationresponseandultimatestrength.Chapter5describesanexperimental
program on short and intermediate length coldformed steel columns with holes.
Chapter 6 initiates the development of a nonlinear finite element protocol with a
significant effort to define the residual stresses and initial plastic strains from the
manufacturing process. The capabilities of the commercial program ABAQUS
(ABAQUS2007a)areexploredatthebeginningofChapter7withpreliminarynonlinear
finite element simulation studies on rectangular plates with holes. The experimental
results from Chapter 5 are then employed in Phase III to fully develop and verify the
modeling protocol. The research culminates in Chapter 8 with the development of a
database of simulated tests which are used in combination with existing experimental
datatovalidatetheDSMdesignmethodforcoldformedsteelmemberswithholes.

Chapter 2
Thin-shell finite element modeling
in ABAQUS
1B

A set of ABAQUS modeling guidelines is formalized in this chapter to provide a

consistentmethodologyforthefiniteelementstudiesconductedinthisthesisresearch.
Finiteelementeigenbucklinganalysisisavaluabletoolforstudyingtheelasticbuckling
properties of thinwalled structures. The accuracy of an analysis is influenced by
decisions made while implementing the finite element model, including the choice of
finiteelementtypeandthemeshinggeometryanddensity.Studiesarepresentedhere
which compare finite element eigenbuckling predictions of plate buckling problems to
known theoretical solutions. The eigenbuckling analyses are performed with the
commercial finite element program ABAQUS (ABAQUS 2007a). The accuracy of
ABAQUSthinshellelementsareevaluated,andfiniteelementconvergencestudiesare
presentedtoidentifylimitsonelementaspectratio.Rulesformodelingroundedcorners
andmeshingaroundholesarealsoprovidedwithsupportingelasticbucklingstudies.

2.1 ComparisonofABAQUSthinshellelements
2B

ThreeABAQUSfiniteelementscommonlyemployedintheelasticbucklinganalysis

ofthinwalledstructuresaretheS9R5,S4,andS4RelementsasshowninFigure2.1.The
83H

S4 and the S4R finite elements are four node general purpose shell elements valid for
both thick and thin shell problems (ABAQUS 2007a). Both elements employ linear
shape functions to interpolate deformation between nodes. The S9R5 element is a
doublycurved thin shell element with nine nodes derived with shear flexible Mindlin
strain definitions and Kirchoff constraints (classical plate theory with no transverse
sheardeformation)enforcedaspenaltyfunctions(Schafer1997).Thiselementemploys
quadraticshapefunctions(resultingfromtheincreaseinthenumberofnodesfrom4to
9)whichprovidetwoimportantbenefitswhenmodelingthinwalledstructures:(1)the
abilitytodefineinitiallycurvedgeometriesand(2)theabilitytoapproximateahalfsine
wave with just one element. The 5 in S9R5 denotes that each element node has 5
degreesoffreedom(threetranslational,tworotational)insteadof6(threetranslational,
three rotational). The rotation of a node about the axis normal to the element mid
surfaceis removedfromtheelementformulationtoimprovecomputationalefficiency.
The R in the S9R5 (and S4R) designation denotes that the calculation of the element
stiffnessisnotexact;thenumberofGaussianintegrationpointsisreducedto improve
computational efficiency and to avoid shear locking. This reduced integration
approach underestimates element stiffness and sometimes results in artificial element
deformation modes with zero strain across the element, commonly referred to as
hourglassmodes(Schafer1997).Theaccuracyofeigenbucklingfiniteelementmodels

arecomparedhereforeachoftheseABAQUSelementtypesagainsttheexactsolutions
fortwocommonplatebucklingproblems.

S9R5

S4/S4R

Figure2.1ABAQUSS4\S4Rshellelementwithfournodesandalinearshapefunction,ABAQUSS9R5shell
elementwithninenodesandaquadraticshapefunction

1.1 Modeling accuracy for a stiffened element


52B

Elastic buckling analyses of a stiffened element were performed in ABAQUS to

compare the accuracy of the ABAQUS S9R5, S4, and S4R elements against a known
solution. A stiffened element is a common term used in thinwalled structures to
describe a crosssectional element restrained on both edges (see Figure 3.1) which is
84H

approximated as a thin simplysupported plate (with sides free to wave) and loaded
uniaxiallyasshowninFigure2.2.
85H

fcr

L
h
Figure2.2Buckledshapeofastiffenedplate

10

Thetheoreticalbucklingstressforastiffenedelementis:
2

2E t
f cr = k
,
12(1 2 ) h

(2.1)

wherehisthewidthoftheplate,Eisthemodulusofelasticityoftheplatematerial, is
thePoissonsratio,andtisthethicknessoftheplate.
Thebucklingcoefficientkis:
2

mh n 2 L
,
k =
+
mh
L

(2.2)

whereListhelengthoftheplateandmandnarethenumberofhalfwavelengthsinthe
longitudinal and transverse directions, respectively (Chajes 1974). In Figure 2.2, m=4
86H

andn=1.

Plate buckling coefficients (k) are approximated in ABAQUS by performing

eigenbucklinganalysesofstiffenedelementswithABAQUSS4,S4R,andS9R5elements.
The element aspect ratio is set at 8:1 for the S9R5 element and 4:1 for the S4R and S4
elementstoensureaconsistentcomparisonbetweenfiniteelementmodels(i.e.,similar
numbersofnodesandcomputationaldemand).Theseparticularelementaspectratios
were also chosen because they are expected to be towards the upper limit of what is
requiredtodiscretizethegeometryofcoldformedsteelmembers(especiallyatrounded
cornerswheretheelementaspectratiocanbequitehigh).Theplatethicknessissetto
t=0.0346 in. E=29500 ksi and =0.30 for all finite element models. The ABAQUS
boundaryandloadingconditionsareimplementedasshowninFigure3.1.
87H

11

Figure 2.3 compares the theoretical k from Eq. (2.1) to the ABAQUS buckling
8H

89H

coefficientsforvaryingplateaspectratios(L/h).TheS9R5elementperformsaccurately
overtherangeofelementaspectratiosconsidered,withamaximumerrorof1.3percent.
TheS4andS4Relementsarenotasaccurate,withmaximumerrorsof11.4percentand
9.7percent,respectively.TheaccuracyoftheplatemodelswithS4andS4Relements
increase with increasing plate aspect ratio, which indirectly implies that solution
accuracy increases as the number of elements per halfwave increase (in the loaded
direction).Thishypothesisisconsistentwiththeelementformulations,sincetheS9R5
element uses a quadratic shape function to estimate displacements (and can therefore
capture the halfsine wave of a buckled plate with as little as one element) and the S4
and S4R elements use linear shape functions (requiring at least three elements to
coarsely estimate the shape of a half sine wave). The S4R element is observed to be
slightlylessstiffthantheS4elementin Figure2.3,whichishypothesizedtooccurasa
890H

resultofthereducedintegrationstiffnessapproximation.

Comparing the number of elements required to model a buckled halfwave is a

moreusefulindicatorofmeshdensityandmodelaccuracythanjusttheelementaspect
ratio alone. Figure 2.4 verifies this supposition by demonstrating the improvement in
891H

modeling accuracy for a stiffened element as the number of finite elements per square
halfwaveincrease.TheS4elementexperiencesmembranelockingwhenthenumberof
elementsperhalfwaveislessthan2,resultinginexceedinglyunconservativevaluesfor
k. The S4R avoids this membrane locking with a reduced integration scheme that
assumes the membrane stiffness is constant in the element (ABAQUS 2007a).

12

Regardless, the accuracy of the S4R element degrades when less than 5 elements per
halfwavelengthareusedandneitherfournodeelement(i.e.,theS4ortheS4R)isableto
capturethesinusoidalshapeofthebuckledhalfwavewithlessthanthreeelementsper
buckled halfwave. The S9R5 accurately predicts the shape of the buckled halfwave
and the buckling coefficient k with just one element. k is within 2.1 percent of the
theoretical value for one element per halfwave and reduces to 0.1 percent for two
elementsperhalfwave.

6
S4
S4R
S9R5
Theory

plate buckling coefficient, k

5.5

S4\S4R (4:1)

S9R5 (8:1)

5
L/h=2 shown

4.5

3.5

0.5

1.5
2
2.5
plate aspect ratio, L/h

3.5

Figure2.3AccuracyofABAQUSS9R5,S4,andS4Relementsforastiffenedelementwithvaryingaspect
ratios,8:1finiteelementaspectratiofortheS9R5element,4:1elementaspectratiofortheS4andS4R
elements

13

7
S4
S4R
S9R5

6.5

buckling coeff., k

6
5.5
5
4.5
4
3.5
3

6
8
10
12
14
16
number of elements per buckled half-wave

18

20

Figure2.4AccuracyofS4,S4R,andS9R5elementsasafunctionofthenumberofelementsprovidedper
buckledhalfwavelength,stiffenedelement,squarewaves(k=4)

1.2 Modeling accuracy for an unstiffened element


53B

An unstiffened element is another common crosssection component considered in

theelasticbucklingofthinwalledcrosssections(see Figure3.1),thebehaviorofwhich
892H

isconservativelyapproximatedasaplatesimplysupportedonthreesidesandfreeon
thefourthsideparalleltothedirectionofauniaxiallyappliedstress.Thebuckledshape
ofanunstiffenedelementisdepictedinFigure2.5.
893H

fcr

L
h

14

Figure2.5Buckledshapeofanunstiffenedelement,m=1shown

Thetheoreticalbucklingcoefficient k ofanunstiffenedelementcanbecalculatedwith
thenumericalsolutionofthefollowingequations(Timoshenko1961):

2
m 2 2
m 2 2
tanh(h ) ,

(
)
tanh

h
=
+

L2
L2

m 2 2 m 2 1 2 2
m 2 2 m 2 1 2 2
k ,and = 2 +
k .
= 2 +
Lh
L
Lh
L

(2.3)

(2.4)

Figure 2.6 compares the theoretical to predicted k versus the number of S9R5
894H

elementsprovidedalongthelengthLofanunstiffenedelement.Theplatedimensions
areheldconstantatL/h=4,whiletheelementaspectratioisvariedfrom1:1to64:1.The
S9R5elementproducesanerrorof4.3percentwithanelementaspectratioof16:1and
anerrorof1.0percentwithanelementaspectratioof8:1.

0.8
0.7

buckling coefficient, k

0.6

8:1

0.5

4:1

1:1

2:1

16:1
0.4

32:1

0.3

element aspect
ratio (typ.)

0.2
0.1
0

k=0.486 when
L/h=4

64:1
0

10

20
30
40
50
number of S9R5 elements along length of plate

60

70

Figure2.6AccuracyofS9R5elementsasthenumberoffiniteelementsprovidedalonganunstiffened
elementvaries,L/h=4

15

2.2 ModelingholesinABAQUS
23B

The ability to incorporate holes into the geometry of an ABAQUS finite element

modelisakeyprerequisitetostudyingtheinfluenceofholesonthestructuralbehavior
ofcoldformedsteelstructuralmembers.Toclearthishurdle,customMatlabcodewas
writtenbytheauthorwhichgeneratesafiniteelementmeshofaplatecontainingahole
(Mathworks2007).Thecodediscretizesthegeometryaroundaholebycreatinglayers
ofS9R5elementsasshownin Figure2.7foraslottedhole,acircularhole,andasquare
895H

hole. (See Appendix A for a description of the custom mesh generation program.
896H

Additional Matlab tools were developed to integrate the hole mesh geometry into an
existing finite element model.) The discretization results in S9R5 elements with
opposite edges which are not initially parallel. The initial geometry of 9 node
quadrilateralelementswithoutparalleledgescanbedefinedwithoutlossofaccuracyas
longasthemidlinenodesremaincenteredbetweenthecornernodes(Cook1989),which
is an advantage over the S4 and S4R elements. ABAQUS recommends that the angle
betweenisoparametriclines(i.e.,corneranglesofanelement)shouldnotbelessthan45
degrees or greater than 135 degrees to ensure accurate numerical integration of the
elementstiffnessmatrix(ABAQUS2007a).Thislimitcoincideswiththeminimumand
maximum S9R5 corner angles for the elements at the bisection of the 90 degree plate
cornersasshowninFigure2.8.
897H

ThisstudyestablishesABAQUSS9R5finiteelementmeshguidelinesforplateswith

holesbystudyingtheconvergenceoftheelasticstabilitysolutionaselementaspectratio

16

varies. Figure 2.7 provides the typical mesh layout and summarizes the plate
89H

dimensions considered in this study. The plate is modeled as a stiffened element in


ABAQUS, simply supported on four sides and loaded uniaxially in compression (see
Figure 3.5 for the ABAQUS implementation of the boundary and loading conditions).
89H

Theplatethicknesst=0.0346in.,E=29500ksi,and=0.30forallfiniteelementmodels.
L=3.4 in.

L=6.0 in.

h=3.4 in.

hhole=1.5 in.

Lhole=4.0 in.

Figure2.7.Finiteelementmeshandplatedimensions:slotted,rectangular,andcircularholes

Theconvergenceoftheelasticbucklingsolutionfortheplateswithholesisstudied

by varying the S9R5 element aspect ratio (a:b) at the bisection of the plate corners as
showninFigure2.8,whereaandbaredefinedas
90H

a=

h hhole
h
, b = hole .
N elem
N layers 2

(2.5)

Theaspectratioisvariedbyincreasingthenumberoffiniteelementlayersaroundthe
hole(Nlayers)whilethemaintainingthenumberofedgeelements(Nelem)constant(i.e.,the
meshdensityincreasesbuttheelementcorneranglesremainconstant).
Nelem =10
45

Nlayers =4

135

Element aspect ratio is a:b


Nlayers number of element layers around hole
Nelem number of edge elements

17

Figure2.8HolediscretizationusingS9R5elements

Figure 2.9 demonstrates that the critical elastic buckling stress for the lowest
901H

bucklingmode(ahalfsinewaveinthiscase)forallholetypesconvergestoaconstant
magnitude when a:b is between 0.5 and 2. This result is employed as a modeling
guidelinefortheresearchworkinthisthesiswiththeexpressionsforaandbinEq.(2.5):
902H

0.5

N elem
N layers

1 2 .
2 hhole

(2.6)

1.2
circular hole
square hole
SSMA slotted hole

fcr,hole/fcr,no hole

0.8

0.6

0.4

0.2

0.5

1.5
2
2.5
S9R5 element aspect ratio, a/b

3.5

Figure2.9ThecriticalelasticbucklingstressconvergestoaconstantmagnitudewhentheS9R5element
aspectratioa/bisbetween0.5and2andelementcorneranglesareskewed

2.3 ModelingRoundedCornersinABAQUS
24B

TheS9R5elementcanbedefinedwithaninitialcurvedgeometryinABAQUSwhich

makesitconvenientformodelingroundedcornersofacoldformedsteelcrosssection.
ABAQUSrecommendsthattheinitialelementcurvatureshouldbelessthan10degrees,

18

where curvature of an S9R5 element is defined as the angle subtended by the nodal
normal and the average element normal as shown in Figure 2.10. The derivation in
903H

Figure2.10demonstratesthatthiscurvaturerecommendationismetwhenfiveormore
904H

S9R5 elements form the 90 degree corner. This limit is unfavorable from a modeling
perspectivebecausetheelementaspectratioincreasesasthenumberofelementsaround
thecornerincrease,anotherpotentialsourceofaccuracydegradation.Also,forafinite
elementmodelwithfour90degreecorners(e.g.,acoldformedsteellippedCsection),
increasing the number of elements at a corner can result in a considerable increase in
computational demand if the corner elements comprise a large proportion of the total
numberofelementsinacrosssection.

Average S9R5
element normal

S = r ,

S
S

Node normal (typ.)

ABAQUS recommends 10 degrees


to limit initial S9R5 element curvature

r
18

18

, S 90 =

N elem

r
2

S 90
, N elem 4.5
2S

Figure2.10ABAQUSS9R5initialcurvaturelimitrequiresatleastfiveelementstomodelcorner

AparameterstudywasconductedtoevaluatetheinfluenceofthenumberofS9R5

elements making up a 90 degree corner on the critical elastic buckling loads for local
buckling(Pcrl),distortionalbuckling(Pcrd),andglobalbuckling(Pcre)ofanSSMA600S162
68Csectioncolumn.Thenumberofcornerelementswerevariedfrom1to5,withthe
associated S9R5 aspect ratio a:b varying from 5 to 22. The column length was held
constant at L=48 inches for all models to accommodate multiple local and distortional
halfwaves.Thecolumnswereloadeduniaxiallyandmodeledwithwarpingfreeends

19

(CUFSMstyleboundaryconditions)asshownin Figure4.2.E=29500ksiand=0.30for
905H

allfiniteelementmodels.Pyisthesquashloadofthecolumncalculatedwiththesteel
yieldstressFy=50ksi.

Figure 2.11 provides the typical mesh geometry of the column and compares a C
906H

sectioncornermodeledasasmoothsurfacewithoneS9R5elementandwiththreeS9R5
elements.Figure 2.12 demonstrates that the number of S9R5 corner elements has a
907H

minimal influence on the elastic buckling behavior of the column, with a slight
decreasingtrend(lessthan1%)incriticalelasticbucklingloadwithincreasingelement
quantity.Meshrefinementatthecornersdoesnotinfluencesolutionaccuracybecause
elastic buckling deformation occurs primarily within the more flexible crosssectional
elements.Ifthesimulationofsharpfoldingofthecornersisrequired,suchasinthecase
of nonlinear finite element modeling to collapse, additional corner elements may be
warrantedtoaccuratelycapturelocalizeddeformationgradients.

Figure2.11SSMA600S16268Csectioncornermodeledwitha)oneS9R5element,b)threeS9R5elements

20

1
0.9
0.8
0.7

Pcr/Py

0.6
0.5
0.4
Global - flexural torsional
Global - weak axis flexure
Distortional
Local

0.3
0.2
0.1
0

2
3
4
Number of corner S9R5 elements

Figure2.12ThenumberofS9R5cornerelementshasaminimalinfluenceonthecriticalelasticbuckling
loadsofanSSMA60016268CsectioncolumnwithL=48in.

2.4 Summaryofmodelingguidelines
25B

TheS9R5elementwillbeimplementedinthisresearchworkbasedonitsversatility

anddemonstratedaccuracy.TheresultsoftheABAQUSstudiesinthischapterformthe
basis of the ABAQUS modeling guidelines below which will be implemented for both
eigenbucklingandnonlinearfiniteelementstudiesinthisthesis:

A minimum of two S9R5 elements per halfwavelength shall be provided in


stiffenedelementsinthedirectionnormaltotheappliedload(e.g.,flangesand
webofalippedCsection)
The S9R5 element aspect ratio shall be less than or equal to 8:1 in unstiffened
elements(e.g.,flangelipinaCsection)
TheS9R5elementaspectratioshallbebetween0.5and2.0whenmodelingholes
withthediscretizationschemedescribedinSection 2.2(wheretheelementsides
arenotperpendicular)
Forbothstiffenedandunstiffenedelements,atleasttwoS9R5elementsshallbe
providedinthedirectionperpendiculartotheapplicationofload
Rounded corners shall be modeled with at least two S9R5 elements, and the
elementaspectratiooftheseelementsshallbelessthanorequalto16:1.
908H

21

Chapter 3
Elastic buckling of cold-formed steel
cross-sectional elements with holes
2B

Asimplifiedmethodfordeterminingtheelasticbucklingpropertiesofathinwalled
crosssection is to evaluate the contribution of each element in the crosssection
separately.Thiselementbyelementevaluationisthebasisoftheeffectivewidthdesign
method for coldformed steel beams and columns and can also be employed as a
conservative predictor of the local critical elastic buckling load (Pcrl) when designing
coldformedsteelmemberswiththeDirectStrengthMethod(AISIS1002007,Appendix
1). The two common crosssection element types in an open thinwalled cross section
are stiffened and unstiffened elements, examples of which are provided in Figure 3.1.
90H

The boundary conditions of a stiffened element are conservatively approximated as a


simplysupportedplate.Theunstiffenedelementistreatedasaplatesimplysupported
onthreesidesandfreeonthefourthedgeparalleltotheapplicationofload.

22

Flange, stiffened
element

Lip,
unstiffened
element

Web,
stiffened
element

Figure3.1StiffenedandunstiffenedelementsinalippedCsection

Theinfluenceofholesontheelasticbucklingbehaviorofstiffenedandunstiffened
elements is evaluated in this chapter using thin shell finite element eigenbuckling
analysis.Thepresenceofholescanmodifythebuckledmodeshapeofanelementand
eitherincreaseordecreaseitscriticalelasticbucklingstress.Holespacingandholesize
relative to element size are studied for both stiffened and unstiffened elements, and
approximate methods for predicting element critical elastic buckling stress are
developedandpresentedforuseindesign.Theresearchresultspresentedherewillbe
usedasaframeworkfortheelasticbucklingstudiesoffullcoldformedsteelstructural
memberswithholesinChapter4.
910H

3.1 Plateandholedimensions
26B

Plateandholedimensionnomenclatureusedthroughoutthischapterissummarized

inFigure3.2.Thestripsofplatebetweenaholeandtheplateedgeswillbereferredtoas
91H

unstiffened strip A and unstiffened strip B, where the widths of these unstiffened
strips are hA and hB respectively as shown in Figure 3.3. For stiffened elements in
912H

bending,theneutralaxislocationisdefinedasYin Figure3.4andismeasuredfromthe
913H

compressededgeoftheplate.

23


S/2

Detail A

Plate with holes


C
L Hole

hhole

Lhole

+hole

C
L Plate
rhole
Slotted hole

Detail A

Figure3.2Elementandholedimensiondefinitions

Unstiffened strip A

hA

hB

Unstiffened strip B
Detail A

Figure3.3DefinitionofunstiffenedstripAandBforaplatewithholes.
Compressed edge

Neutral axis
Y

Tension edge
Detail A

Figure3.4Definitionofneutralaxislocationforstiffenedelementsinbending.

24

3.2 Finiteelementmodelingassumptions
27B

The elastic buckling behavior of stiffened and unstiffened elements with holes are

obtained with eigenbuckling analyses of plates in ABAQUS (ABAQUS 2007a). All


members are modeled with ABAQUS S9R5 reduced integration ninenode thin shell
elements.Thetypicalfiniteelementaspectratiois1:1andthemaximumaspectratiois
limitedto8:1(referto Chapter2foradiscussiononABAQUSthinshellfiniteelement
914H

types and finite element aspect ratio limits). Element meshing was performed with a
Matlab (Mathworks 2007) program written by the author (refer to Appendix A for a
915H

description of the program). The plate models are loaded from each end with stress
distributions applied as consistent nodal loads in ABAQUS. Converting a stress
distribution to consistent nodal loads for the S9R5 element requires a different
procedure than that followed for a 4node finite element (Schafer 1997). Coldformed
steel material properties are assumed as E=29500 ksi and =0.3 in all finite element
models.

3.3 Stiffenedelementinuniaxialcompression
28B

3.1 Boundary and loading conditions


54B

Thestiffenedelementismodeledwithsimplysupportedboundaryconditionsand
loadeduniaxiallywithauniformcompressivestressasshowninFigure3.5.
916H

25

perimeter
supported in 2 (v = 0)
longitudinal midline
supported in 1 (u = 0)
2

transverse midline
supported in 3 (w = 0)

Figure3.5ABAQUSboundaryconditionsandloadingconditionsforastiffenedelementinuniaxial
compression

3.2 Influence of a single slotted hole


5B

This study explores the influence of a single slotted hole on the elastic buckling
stress of a stiffened element. The plate length L is varied from three to twentyfour
timestheslottedholelength,Lhole,andthewidthoftheplatesarechosentoequaltheflat
web widths of four common Steel Stud Manufacturers Association (SSMA) structural
studs listed in Table 3.1 (SSMA 2001). The slotted hole has dimensions of hhole=1.5 in.,
917H

Lhole=4in.,andrhole=0.75in.Holesarealwayscenteredtransverselybetweentheunloaded
edgesoftheplateinthisstudy.Theplatethickness,t,is0.0346in.

Table3.1PlatewidthscorrespondingtoSSMAstructuralstuddesignations

SSMA

Designation

(in)

250S162-33

2.28

0.66

362S162-33

3.40

0.44

600S162-33

5.78

0.26

800S162-33

7.78

0.19

hhole/h

The results of this study are presented in Figure 3.6 and demonstrate that as the
918H

length of a stiffened element increases relative to the length of the hole, the critical
elasticbucklingstress,fcr,convergestoaconstantmagnitudewhichiseitherequaltoor
lowerthanthebucklingstressofaplatewithoutahole.TheconvergenceoccursasL/Lhole

26

exceeds 5, suggesting that the influence of the hole is independent of the plate end
conditionsbeyondthislength.

1.2

fcr,hole/fcr,no hole

1
0.8
hhole/h=0.66
hhole/h=0.44

0.6

hhole/h=0.19
hhole/h=0.26

0.4

0.2

10

15

20

25

L/Lhole

Figure3.6Influenceofaslottedholeontheelasticbucklingstressofasimplysupportedrectangularplate
withvaryinglength

Whentheholeiswiderelativetothewidthoftheplate(hhole/h=0.66)andL/Lholeis

small(seeFigure3.6),theelasticbucklingstressoftheplatewiththeholeisasmuchas7
91H

percenthigherthanforaplatewithoutahole.Thisincreaseinstressisexplainedbythe
buckledmodeshapesin Figure3.7.Theplatewiththeholein Figure3.7ahasahigher
920H

921H

elasticbucklingstressthantheplatewithouttheholein Figure3.7bbecausethenatural
92H

patternofbuckledwavesismodifiedbythehole.Thebuckledcellsadjacenttothehole
areshorterandthereforestiffer.Thethinstripsattheholedampenbucklinginthiscase
becausetheyhaveanaxialstiffnesshigherthanthebuckledcellsawayfromthehole.

27

(a)

(b)

Figure3.7Comparisonofbuckledshapeanddisplacementcontoursforarectangularplatewithhhole/h=0.66
andL/Lhole=3,(a)withslottedholeand(b)withouthole.Noticethechangeinlengthandquantityofbuckled
cellswiththeadditionofaslottedhole.

As the plate length increases past L/Lhole=5 for the smallest plate width
(hhole/h=0.66),thebucklingstressconvergestothatofaplatewithoutahole. Figure3.8
923H

demonstrates that for these long, slender stiffened elements the slotted hole dampens
bucklingneartheholebutdoesnotappreciablychangethenaturalhalfwavelengthof
thebuckledcellsaswasobservedfortheshorterplatesinFigure3.7.
924H

(a)

(b)

Figure3.8Buckledshapeofasimplysupportedplate(a)withaslottedholeand(b)withoutahole.
L=15Lhole,hhole/h=0.66.Theslottedholedampensbucklingbutdoesnotsignificantlychangethenaturalhalf
wavelengthoftheplate.

For plates with hhole/h less than 0.66, the slotted hole causes a decrease in the
elastic buckling stress which converges to a constant magnitude as the plate length
exceedsL/Lhole=5. Figure3.9ademonstratesthatlocalbucklingneartheholecontrolsthe
925H

28

elasticbucklingstressofthesewiderplates.Thedeformationattheholeresultsfromthe
localizedreductionintransverseplatebendingstiffness.
AsplatelengthdecreasesbelowL/Lhole<5andhhole/h=0.19,theinfluenceofthehole
onthecriticalelasticbucklingstressfluctuatesasshowninFigure3.6.Whenthelowest
926H

elastic buckling mode shape results in an odd number of halfwaves, the hole falls
within the central halfwave and the critical elastic buckling stress decreases. For an
evennumberofhalfwaves,theholeislocatedatthetransitionbetweentwohalfsine
waves (because the hole is centeredat the midlength of theplate), forcing the buckled
cellstoshortenandincreasingthecriticalelasticbucklingstress.

(a)

(b)

Figure3.9(a)Slottedholecauseslocalbuckling(hhole/h=0.26),comparedto(b)buckledcellsatthenatural
halfwavelengthoftheplate

3.3 Influence of slotted hole spacing


56B

Thepreviousstudydemonstratedthattheelasticbucklingbehaviorofastiffened
element with a single hole is sensitive to the size of the hole relative to the size of the
plate. The focus now shifts to the influence of multiple slotted holes on the elastic
bucklingstressofalongfixedlengthstiffenedelement.Inthisstudy,slottedholesare
addedonebyonetoastiffenedelement(whereL=24Lhole)suchthatthecentertocenter
spacingSvariesasshowninFigure3.10.
927H

29

Lhole

S/2

hhole

Figure3.10Definitionofcentertocenterdimensionfortheslottedholes

Asholespacingdecreases,theelasticbucklingstressinFigure3.11eitherincreases
928H

ordecreasesdependingontheratioofholewidthtoplatewidth.Whentherearemany
large holes (hhole/h=0.66, S/ Lhole < 4), buckling is dampened at the holes and the buckled
cells shorten their lengths to form between adjacent holes (see Figure 3.12 for buckled
92H

shape).Thedecreaseinbuckledhalfwavelengthcausesanincreaseinelasticbuckling
stressoftheplate.

1.2

fcr,holes /fcr,no holes

0.8

0.6

0.4

0.9

hhole/h=0.66

0.85

hhole/h=0.44

0.75
0.2

hhole/h=0.19

0.8

hhole/h=0.26
2

3
4
S/Lhole

10

15
S/Lhole

20

25

Figure3.11Influenceofslottedholespacingontheelasticbucklingloadofalongsimplysupported
rectangularplate

30

Whentheholesaresmallerrelativetotheplatewidth(hhole/h<0.44)andarespaced
closely together (S/ Lhole < 4), the local buckling influence of adjacent holes combine to
sharply decrease the elastic buckling stress. The inset of Figure 3.11 highlights this
930H

reduction in elastic buckling stress for hhole/h=0.19 and hhole/h=0.26, and Figure 3.12
931H

provides a summary of the associated buckled shapes. When hole spacing increases
beyondS/Lhole=5,theelasticbucklingstressesapproachconstantmagnitudesforallplate
widths considered, which is consistent with the trends presented in Figure 3.6. This
932H

observationisimportantfromadesignperspectivebecauseitservesasarationalbasis
forsettingholespacinglimitsincoldformedsteelmembers.

Buckling is dampened at the holes, halfwaves form between holes

hhole/h=0.66, S/Lhole=4
Buckling of the unstiffened strips adjacent to
the hole is dominant here

hhole/h=0.44, S/Lhole=4
Buckled half-waves form along the length of
the plate

hhole/h=0.26, S/Lhole=4

Figure3.12Comparisonofbuckledshapesforalongstiffenedelement(L=24Lhole)withaslottedhole
spacingofS/Lhole=4andhhole/h=0.66,0.44,and0.26.

Figure3.12highlightsthetwotypesofbucklingmodesthatcanoccurinstiffened
93H

elements,platebucklingandunstiffenedstripbuckling.Theinfluenceofthesebuckling

31

modes on fcr is reflected in Figure 3.13. The maximum decrease in fcr occurs for a
934H

relatively small hole when compared to the plate width (hhole/h=0.30) and lies at the
transition between plate buckling, where axial stiffness of the buckled cells is reduced
with the presence of holes, and unstiffened strip buckling. Unstiffened strip buckling
occursbetweenhhole/h=0.30andhhole/h=0.55resultinginarelativeincreaseinfcrasthestrips
adjacent to the holes increase the axial stiffness of the plate. As hhole/h increases past
hhole/h=0.55theunstiffenedstripadjacenttotheholebecomesnarrowandstiff,resulting
inplatebucklingawayfromtheholesandanfcrsimilartoaplatewithoutahole.(An
increaseincriticalelasticbucklingstressforlargeholesdoesnotnecessarilycorrespond
toanincreaseinultimatestrengthbecausethestrengthoftheplatewillbelimitedbythe
strength of the net crosssection.) This is another important observation that will be
usedwhendevelopinganelasticbucklingpredictionmethodforstiffenedelementswith
holesinSection3.3.4.
935H

32

fcr,holes /fcr,no holes

1.5

Plate buckling

0.5

0.2

Unstiffened strip
buckling

0.4

Plate buckling
away from the hole

0.6
hhole/h

0.8

Figure3.13Variationinfcrwithincreasinghhole/hforastiffenedelementcorrespondtobucklingmodeshapes
(seeFigure3.12forexamplesofplatebucklingandunstiffenedstripbucklingmodeshapes)

936H

3.4 Approximate prediction method for use in design


57B

Approximations for the critical elastic buckling stress of stiffened elements (e.g.

columnweborflangeofalippedCsection)withholesunderuniaxialcompressionare
developed in this section considering two elastic buckling states, buckling of the plate
withoutholeinfluenceandbucklingoftheunstiffenedstripsadjacenttothehole.The
proposed prediction method is validated with thin shell finite element eigenbuckling
analyses for a variety of hole shapes, sizes, and spacings. Mandatory dimensional
tolerances on the prediction method are explicitly defined, and optional dimensional
limits,markedwithanasterisk(*),areprovidedtoavoidexcessiveconservatism.

33

3.4.1 Definitionsandassumptions
18B

Figure 3.2 defines the plate and hole dimension notation used in the element
937H

predictionmethod,includingtheholespacingS,platewidthh,andholelengthandhole
width,Lholeandhhole. holeistheoffsetdistanceofaholemeasuredfromthecenterlineof
the plate. The elastic buckling prediction method for a stiffened element is developed
assuming a long plate loaded uniaxially and simplysupported on all four sides with
evenly spaced holes. A summary of all prediction method equations is provided in
AppendixD.
938H

3.4.2 PredictionEquations
19B

Theelasticbucklingstressofastiffenedelementwithholesisapproximatedas

f crl = min[ f cr , f crh ].

(3.1)

Thecriticalelasticbucklingstressforplatebuckling(withoutholeinfluence)is
2

2E t
f cr = k
,
12(1 2 ) h

(3.2)

wherekiscommonlytakenequalto4whenconsideringlongrectangularplates(L/h>4).
When elastic buckling of the stiffened element is governed by the buckling of an
unstiffenedstripadjacenttothehole,thecriticalelasticbucklingstressofthegoverning
unstiffenedstripis:

f crh ,net = min [ f crA , f crB ]

34

(3.3)

f cri

2E t

= ki
12(1 2 ) hi

andi=AorB

(3.4)

TheplatebucklingcoefficientkiforunstiffenedstripsAandBareapproximatedby(Yu
andSchafer2007):

for Lhole hi 1 ,

for Lhole hi < 1 ,

k i = 0.425 +

(Lhole

0.2
,
0.95
hi ) 0.6

(3.5)

ki = 0.925 ,andi=AorB.

(3.6)

Eq.(3.5)accountsforthelengthoftheunstiffenedstrip.Asholelengthshortensrelative
93H

totheunstiffenedstripwidth,kiincreases.ThisisanimprovementoverAISIS100which
conservativelyassumesthelowerboundk=0.425regardlessofholelength.WhenLhole/hiis
less than 1, k may be conservatively assumed equal to 0.925 via Eq. (3.6) or calculated
940H

directly by solving the classical stability equations for an unstiffened element


(Timoshenko1961).
Anet = (h hhole ) t

Ag = ht

P2
P1 + P2 = Pcr = f crh ,net Anet

hhole

Pcr = f crh Ag

P1
fcrh,net

fcrh,net is the critical elastic buckling


stress of the wider unstiffened strip

f crh = f crh ,net

Anet
h
= f crh ,net 1 hole
Ag
h

Figure3.14Unstiffenedstripelasticbucklingstressconversionfromthenettothegrosssection

35

Tocomparethebucklingstressfromtheunstiffenedstrip(fcrh,net)tothatoftheentireplate
(fcr) equilibrium between the net and gross section must be considered, as shown in
Figure3.14andprovidedinthefollowing:
941H

f crh = f crh ,net (1 hhole h ) .

(3.7)

3.4.3 Verificationandequationlimits
120B

3.3.4.3.1

Holescenteredtransverselyinplate
184B

ThinshellfiniteelementeigenbucklinganalysisinABAQUS,asdescribedinSection

3.2, is employed here to verify the accuracy of the approximate prediction method in
942H

Section 3.3.4.2.Theboundaryandloadingconditionsassumedforthestiffenedelement
943H

aredescribedin Figure3.5.Thelengthoftheslottedhole,Lhole,widthoftheplateh,the
94H

shape of hole (slotted, circular, square), the hole spacing S, length of the plate L, and
platethicknesstarevariedintheanalyses.Theplateandholedimensionsaswellasthe
ABAQUScriticalelasticbucklingstress,fcrl,forthe145modelsconsidered,areprovided
in Appendix B(the eigenbuckling results from the studies in Section 3.3.2 and Section
945H

946H

3.3.3 are included in the 145 models). The parametric ranges in this study are
947H

summarizedforeachholetypeinTable3.2.
948H

Table3.2Parameterrangesinstiffenedelementverificationstudy.
Hole type
Slotted
Circular
Square

Min
Max
Min
Max
Min
Max

hhole/h
0.10
0.70
0.10
0.70
0.10
0.70

S/Lhole
1.7
24.0
13.3
13.3
13.3
13.3

36

S/h
1.2
42.2
1.3
9.3
1.3
9.3

h/t
21
434
62
434
62
434

# of models
131
7
7

The results of the ABAQUS eigenbuckling analyses are compared to the stiffened
elementpredictionmethodinFigure3.15andFigure3.16.Figure3.15demonstratesthat
94H

950H

951H

asholespacingSbecomessmallrelativetotheplatewidthh,thepredictionmethodis
not alwaysaccurate. As hole spacing decreases, holes begin to coincide with thelocal
bucklinghalfwavelengths(whichhavealengthofh)andtheinfluenceoftheindividual
holes act cumulatively to decrease the axial stiffness of the plate. A similar loss in
stiffness is observed in Figure 3.16 as hole spacing decreases relative to hole length.
952H

Fromtheseobservations,thefollowinglimitsareimposedonthepredictionmethod:

S
1.5 ,
h

(3.8)

S
2 .
Lhole

(3.9)

IftheparameterlimitinEq.(3.9)issubstitutedintoEq.(3.8),athirddimensionallimitis
953H

954H

automaticallyimposed:

Lhole
0.75
h

(3.10)

Eq.(3.10)preventstheholelengthfrombeingtoolongrelativetothehalfwavelengthof
95H

theplate.ThemeanandstandarddeviationoftheABAQUStopredictedratioforthe
stiffened elements within the limits of Eq. (3.8) and Eq. (3.9) are 1.02 and 0.04
956H

957H

respectively,demonstratingthatthepredictionmethodisviable.

37

1.5
Plate buckling controls

Plate buckling controls

Unstiffened strip controls

Unstiffened strip controls

fcrl, ABAQUS/fcrl, predicted

fcrl, ABAQUS/fcrl, predicted

1.5

0.5

10

20

30

40

0.5

50

S/h

10

20

30

40

50

S/h

Figure3.15AccuracyofstiffenedelementpredictionmethodasafunctionofholespacingStoplatewidthh
(a)withoutand(b)withthedimensionallimitsinEq.(3.8)andEq.(3.9)
958H

1.5

95H

1.5

Plate buckling controls

Plate buckling controls


Unstiffened strip controls

fcrl, ABAQUS/fcrl, predicted

fcrl, ABAQUS/fcrl, predicted

Unstiffened strip controls

0.5

10

15

20

0.5

25

S/Lhole

10

15

20

25

S/Lhole

Figure3.16AccuracyofstiffenedelementpredictionmethodasafunctionofholespacingStolengthof
holeLhole(a)withoutand(b)withthedimensionallimitsinEq.(3.8)andEq.(3.9)

960H

1.5

1.5
Plate buckling controls
Unstiffened strip controls
fcrl, ABAQUS/fcrl, predicted

fcrl, ABAQUS/fcrl, predicted

Plate buckling controls


Unstiffened strip controls

0.5

0.2

0.4

0.6

0.8

961H

0.5

hhole/h

0.2

0.4

0.6

0.8

hhole/h

Figure3.17Accuracyofthestiffenedelementpredictionmethodasafunctionofholewidthhholetoplate
widthh(a)withoutand(b)withthedimensionallimitsinEq.(3.8)andEq.(3.9)
962H

38

963H

AsholewidthincreasesrelativetoplatewidthinFigure3.17,thecontrollingbuckled
964H

state transitions from buckling of the unstiffened strip adjacent to plate buckling. The
stripsofwebmaterialadjacenttotheholeshaveahigheraxialstiffnessthanthesections
oftheplatewithoutholes,causingplatebucklingtooccurbetweentheholesasshown
inFigure3.18.
965H

Strips of plate adjacent to the


hole are stiffer than plate
between holes

Figure3.18Forplateswheretheunstiffenedstripisnarrowcomparedtotheplatewidth,platebuckling
occursbetweentheholes.

As the hole width becomes small relative to plate width, the unstiffened strip

buckledstateispredictedbythesimplifiedmethodforslottedholes,althoughtheactual
behaviorisacombinationofplatebucklingandlocalbucklingattheholes,asshownin
Figure3.19.Theassumptionofunstiffenedstripbucklingwhentheslottedholewidthis
96H

small relative to plate width is conservative, with a maximum ABAQUS to predicted


ratioof1.16whenhhole/hisintherangeof0.30. Figure3.19alsodemonstratesthatplate
967H

buckling dominates over unstiffened strip buckling for stiffened elements with square
andcircularholes.Thepredictionmethodidentifiesthiselasticbucklingbehaviorand
accurately predicts fcrl as shown in Figure 3.20, where stiffened element results
968H

containingjustsquareorjustcircularholesareplotted.

39


Plate buckling dominates over
unstiffened strip buckling for
square (and circular holes)

Plate buckling and


unstiffened strip buckling
are both present when
0.20hhole/h0.60

Figure3.19Platebucklingandunstiffenedstripbucklingmaybothexistforaplatewithholes.These
modesarepredictedconservativelyasunstiffenedstripbuckling.

1.5

fcr , ABAQUS/fcr , predicted


l
l

Plate buckling controls

0.5

0.2

0.4

0.6

0.8

hhole/h

Figure3.20Accuracyofpredictionmethodforstiffenedelementswithsquareorcircularholesasafunction
ofholewidthhholetoplatewidthh.

3.3.4.3.2

Offsetholes
185B

43 additional ABAQUS eigenbuckling analyses were performed to evaluate the

accuracy of the simplified prediction method in Section 3.3.4.2, but now with
96H

transversely offset holes. For these models, the hole offset from the centerline of the

40

plate, hole, and the plate width, h, were varied. All plate models in this study have
regularly spaced slotted holes (S=20 in.) and constant plate length, L, of 100 in. The
boundary and loading conditions assumed for the stiffened element are described in
Figure 3.5. The model dimensions and critical elastic buckling stress, fcrl, for the 43
970H

modelsconsidered,aresummarizedin AppendixB.hstripisthewidestunstiffenedstrip,
971H

eitherhAandhB.TheparametricrangesforthisstudyaresummarizedinTable3.3.
972H

Table3.3Parameterrangeforstiffenedelementverificationstudywithoffsetholes.
hhole/h

S/Lhole

S/h

h/t

hole/h

Min

0.10

5.0

1.3

62

0.000

Max

0.70

5.0

9.3

434

0.375

Hole type
Slotted

Plate buckling controls

Unstiffened strip controls

Unstiffened strip controls

2.5
fcrl, ABAQUS/fcrl, predicted

2.5
fcrl, ABAQUS/fcrl, predicted

43

3
Plate buckling controls

1.5

0.5

# of models

1.5

0.5

0.2

0.4

0.6

0.8

0.2

hstrip/h

0.4

0.6

0.8

hstrip/h

Figure3.21Accuracyofthestiffenedelementelasticbucklingpredictionmethodasafunctionof
unstiffenedstripwidthhstripversusplatewidthhforoffsetholes(a)withoutand(b)withthedimensional
limitsinEq.(3.8)andEq.(3.9)

973H

974H

The ABAQUS critical elastic buckling stress results are compared to the prediction
methodin Figure3.21,anddemonstratethatthepredictionmethodisconservativeand
975H

thattheaccuracyofthemethodimprovesashstripdecreasesrelativetotheplatewidthh
andholelengthLhole.Theunstiffenedstripbuckledstateispredictedtocontrolformost
of the plate models, primarily because the shift in hole location results in a wider

41

unstiffened strip with less axial stiffness than that provided by the plate material
betweenholes.Whentheplateisrelativelywidecomparedtothewidthoftheholeand
theholeisshiftedneartheedgeoftheplateasshowninFigure3.22,thepredictionscan
976H

be very conservative. The wide unstiffened strip is not a good approximation of the
actualbehavioroftheplateinthiscase.Predictionaccuracyvarieswithholeoffset, hole,
as shown in Figure 3.23a, and is most conservative as the hole offset becomes large
97H

relativetotheplatewidthh.Toavoidoverlyconservativeresults,thefollowinglimiton
holeoffsetholeisproposedforstiffenedelements:

hole

0.15 *

(3.11)

ThemeanandstandarddeviationoftheABAQUStopredictedratioforthedatawithin
thedimensionallimitsofEq. (3.8),Eq. (3.9),andEq. (3.11)are1.14and0.15respectively
978H

97H

980H

(alsoseeFigure3.23b).
981H

The prediction method conservatively assumes


unstiffened strip buckling of the wider strip adjacent to the
hole, although plate buckling is observed.

hstrip

Figure3.22Holesattheedgeofawidestiffenedplatereducetheaxialstiffness(andcriticalelasticbuckling
stress)butdonotchangethebuckledshape.

42

Plate buckling controls

Plate buckling controls


Unstiffened strip controls

1.5

1.5

0.5

0.5

Unstiffened strip controls

2.5
fcr , ABAQUS/fcr , predicted
l
l

fcrl, ABAQUS/fcrl, predicted

2.5

0.1

0.2

0.3

0.4

hole/h

0.5

0.1

0.2

0.3

0.4

0.5

hole/h

Figure3.23Accuracyofthestiffenedelementelasticbucklingpredictionmethodasafunctionofholeoffset

holeversusplatewidthhforoffsetholes(a)withoutand(b)withthedimensionallimitsinEq.(3.8),Eq.(3.9),
982H

983H

andEq.(3.11)

984H

3.4 Stiffenedelementinbending
29B

4.1 Boundary and loading conditions


58B

Thestiffenedelementismodeledwithsimplysupportedboundaryconditionsand
loaded with a bending compressive stress distribution as shown in Figure 3.24. The
985H

location of the neutral axis about which bending occurs, Y, is measured from the
compressededgeoftheplate.
Restrain point in 3
(w=0)
Neutral axis

Restrain transverse
midline in 1 (u=0)

Restrain plate
perimeter in 2 (v=0)
Restrain point in 3
(w=0)

2
1

Figure3.24Boundaryandloadingconditionsforastiffenedelementinbending

43

4.2 Influence of transversely-centered slotted holes


59B

Shell finite element eigenbuckling models of stiffened elements with regularly

spaced slotted holes are evaluated in this study. The bending stress distribution is
symmetric about the transverse centerline of the plate (Y=0.50h) for all models. The
slotted holes are centered transversely in the plate (hole=0). The plate and hole
dimensionsandthecriticalelasticbucklingstress,fcrl,forthe28modelsconsidered,are
summarizedin AppendixB.Theparametricrangesforthisstudyaresummarizedin
986H

Table3.4.
987H

Table3.4Parameterrangesconsideredforstiffenedelementsinbendingwithholes.
Hole type
Slotted

hhole/h

S/Lhole

S/h

h/t

Y/h

Min

0.10

1.67

1.33

61.93

0.50

Max

0.70

5.00

9.33

433.53

0.50

# of models
28

Figure 3.25 highlights the influence of hole width to plate width on stiffened
98H

elements in bending. As hhole/h increases, the buckling mode transitions from plate
buckling (similar to a plate without a hole) to buckling of the compressed unstiffened
stripadjacenttothehole.Thebuckledhalfwavelengthofaplateinbendingisbetween
0.25h to 0.50h, which results in a shortened halfwavelength of the unstiffened strip
(often less than the length of the hole) when compared to the equivalent unstiffened
stripbucklingmodeforstiffenedelementsinuniaxialcompression(SeeSection3.3).
98H

44

Unstiffened strip half-wavelength can


be less than the length of the hole.

hhole/h=0.10

hhole/h=0.30

hhole/h=0.50

Unstiffened strip buckling becomes more


predominant as the hole size increases
relative to plate width.

Figure3.25Stiffenedplatesloadedwithalinearbendingstressgradientexhibitbucklingoftheunstiffened
stripadjacenttotheholeinthecompressionregionoftheplate.

The maximum reduction in critical elastic buckling stress occurs in the range of

hhole/h=0.30 as shown in Figure 3.26a. This result is consistent with the elastic buckling
90H

results for stiffened plates under axial compression (See Figure 3.13). The elastic
91H

buckling behavior of stiffened elements in bending are different than in pure


compression though as hhole/h exceeds 0.50. Unstiffened strip buckling continues to
dominate for plate bending (with an associated reduction in fcr) while plate buckling
away from the hole controls for uniaxially compressed plates (with minimal influence
on fcr even for very large holes). This distinction between compression (columns) and
bending (beams) elastic buckling behavior of stiffened elements is important when
consideringhowtoapproximateelasticbucklingbehavior.fcrdecreasesasholespacing
becomes small relative to hole length as shown in Figure 3.26b, identifying S/Lhole as
92H

another important parameter when predicting elastic buckling of stiffened elements in


bending(asitisforstiffenedelementsincompression,SeeSection3.3).
93H

45

1.5

fcr,hole/fcr,no hole

fcr,hole/fcr,no hole

1.5

0.5

0.2

0.4

0.6

0.8

0.5

hhole/h

10

S/Lhole

Figure3.26Influenceofslottedholesoncriticalelasticbucklingstressfcrofstiffenedelementsinbendingas
afunctionof(a)holesizerelativetoplatewidthand(b)holespacingasafunctionofholelength.

4.3 Influence of offset slotted holes


60B

4.3.1 NeutralaxislocationatY=0.50h
12B

Shell finite element eigenbuckling models of stiffened elements with regularly

spacedoffsetslottedholesareevaluatedinthisstudy.Thebendingstressdistributionis
symmetricaboutthetransversecenterlineoftheplate(Y=0.50h)forallmodels.Thehole
offset,hole,rangesfrom0.375hto+0.375h,whereapositiveshiftmovestheholesintothe
compression region of the plate. The plate and hole dimensions and the critical elastic
bucklingstress,fcr,forthe92modelsconsidered,aresummarizedin AppendixB.The
94H

parameterrangeconsideredinthisstudyisprovidedinTable3.5.
95H

Table3.5Studyparameterlimitsforstiffenedelementinbending(Y/h=0.50)withoffsetholes
hhole/h

S/Lhole

S/h

h/t

Y/h

hole/h

Min

0.10

5.00

1.33

61.93

0.50

-0.375

Max

0.70

5.00

9.33

433.53

0.50

0.375

Hole type
Slotted

# of models
92

The presence of holes in the compression region of a stiffened element in bending

(Y=h/2) decreases the critical elastic buckling stress when compared to a plate without
holesasshownin Figure3.27.DependinguponthewidthoftheunstiffenedstripA
96H

46

inthecompressedregionoftheplateandtheunstiffenedstripBinthetensileregion
of the plate (see Figure 3.3 for definitions) relative to hole depth h, unstiffened strip
97H

buckling may occur above the hole, below the hole, or above and below the hole. fcr
varieswiththetransversepositionoftheholesintheplate(characterizedasthewidthof
unstiffened strip A, hA) in Figure 3.27. The trends in fcr can be related to the elastic
98H

buckling modes in Figure 3.28. If the holes are located in the tensile region of the
9H

stiffenedelement,thebuckledmodeshape(andfcr)areunchangedwhencomparedtoa
stiffenedelementwithoutholes.Therelationshipbetweenthesebuckledmodeshapes
and trends in fcr will be used in Section 3.4.4 when developing an approximate elastic
10H

bucklingpredictionmethodforstiffenedelementsinbending.

0
0.1

hhole/h=0.10

hhole/h=0.20
B

0.2
0.3

hA /h

hhole/h=0.40

Compression

0.4

hhole/h=0.30
hhole/h=0.50

Neutral axis

hhole/h=0.60
hhole/h=0.70

0.5
0.6
D

0.7

Tension

0.8
0.9
1

0.5

1.5

fcr,hole/fcr,no hole

Figure3.27Holelocationinfluenceoncriticalelasticbucklingstressfcrforastiffenedplateinbending
(Y=0.50h)(BuckledmodeshapescorrespondingtoA,B,C,andDareprovidedinFigure3.28.)
10H

47

hA
Unstiffened strip
buckling (below hole)

B
Unstiffened strip
buckling (below and
above hole)

D
Unstiffened strip
buckling (above hole)

Plate buckling (no hole


influence)

Figure3.28Thebuckledmodeshapechangesasslottedholesmovefromthecompressionregiontothe
tensionregionofastiffenedelementinbending(hhole/h=0.20).

4.3.2 NeutralaxislocationatY=0.75h
12B

TheneutralaxisintheshellfiniteelementeigenbucklingmodelsfromSection3.4.3.2
102H

isnowmodifiedtoY=0.75h.ThetrendsinfcrinFigure3.29aresimilartothoseobserved
103H

in Figure3.27(Y=0.50h).Elasticbucklingoftheunstiffenedstripbelowtheholesoccurs
104H

when the hole is close to the compressed edge. The mode shape transitions to
unstiffenedstripbucklingabovetheholesastheholeoffsetincreasestowardthetensile
region of the plate. The plate and hole dimensions and the critical elastic buckling
stress,fcr,forthe92modelsconsideredhere,aresummarizedinAppendixB.
105H

48

0
hhole/h=0.10

0.1

hhole/h=0.20

0.2

hhole/h=0.30

0.3

hhole/h=0.40
hhole/h=0.50

0.4

hhole/h=0.60

hA /h

Compression

hhole/h=0.70

0.5
0.6

Neutral axis

0.7
0.8
Tension

0.9
1

0.5

1
fcr,hole/fcr,no hole

1.5

Figure3.29Holelocationinfluenceoncriticalelasticbucklingstressfcrforastiffenedplateinbending
(Y=0.75h)

4.4 Approximate prediction method for use in design


61B

Intheprevioussectionuniqueelasticbucklingmodeswereidentifiedforastiffened

elementinbendingwithholes.Bucklingoftheunstiffenedstripbetweentheholeand
the compressed edge of the plate (unstiffened strip A) or between the hole and the
tension edge of the plate (unstiffened strip B) may occur depending upon the
transverse location of the hole in the plate, the width of the hole (hhole) relative to the
depth of the plate (h), and the location of the plate neutral axis (Y). If the hole is
completelycontainedwithinthetensionregionoftheplatethentheholehasaminimal
influence on elastic buckling and the critical elastic buckling stress, fcr, remains
unchanged.Theseobservationscanbeusedtodefineanapproximationforthecritical
elasticbucklingstressofastiffenedelementwithholesinbending:

49

f crl = min[ f cr , f crh ] .

(3.12)

The critical elastic buckling stress for a stiffened element in bending (without the
influenceofholes),fcr,maybedeterminedwithEq. (3.2)wherethebucklingcoefficientk
106H

iscalculatedwithAISIS10007Eq.B2.32(AISIS1002007):

k = 4 + 2(1 + ) + 2(1 + )
3

(3.13)

and istheabsolutevalueoftheratiooftensilestresstocompressivestressappliedto
thestiffenedelement,i.e.:

= f 2 f1 = (h Y ) Y .

(3.14)

When elastic buckling of the stiffened element is governed by the buckling of an


unstiffenedstripadjacenttoahole,thecriticalelasticbucklingstressis:

f crh ,net = min [ f crA , f crB ]

(3.15)

ConsiderationofunstiffenedstripAisrequiredonlyifhA<Y,i.e.,atleastaportionof
theholemustlieinthecompressionregionofthestiffenedelement.Ifthatconditionis
mettheelasticbucklingstressforstripAis:
2

f crA

2E t

= kA
12(1 2 ) h A

(3.16)

TheplatebucklingcoefficientfortheunstiffenedstripAisapproximatedas

kA =

2.70 1.76 A
Y hA
0.578
+
,and A =

2
A + 0.34 0.024 A + 0.035 + (Lhole h A )
Y

(3.17)

Eq. (3.17) is a modification of AISIS10007 Eq. B3.32 (AISIS100 2007)This expression


107H

accounts for the gradient of the compressive stress distribution and the aspect ratio of

50

theunstiffenedstrip(see AppendixCforderivation).Theequationfor Aisderivedin


108H

Figure3.30.
109H

hA
f1

Neutral Axis

f2

Similar
Triangles

A =

f 2 Y hA
=
f1
Y

Figure3.30DerivationofstressratioforunstiffenedstripA.

Consideration of unstiffened strip B is required only if hA+hhole<Y, i.e., only when

theentireholelieswithinthecompressedregionoftheplate.Forthiscasethebuckling
stressoftheunstiffenedstrip,convertedtoastressatthecompressededge,isfoundas:

f crB

2E t

= kB
12(1 2 ) hB

Y h A hhole

(3.18)

where the final term in Eq. (3.18) converts the buckling stress from the edge of
10H

unstiffenedstripBtotheedgeofunstiffenedstripAasshowninFigure3.31sothat
10H

thetwostresses(fcrAandfcrB)maybecomparedinEq. (3.15)todeterminetheminimum.
102H

TheplatebucklingcoefficientfortheunstiffenedstripBisapproximatedas:

forLhole/hB>2

forLhole/hB2

k B = 0.340 B2 + 0.100 B + 0.573 ,

51

(3.19)

h
0.38 B + 1.6 B + 0.49
Lhole
,
kB =
0.1

h
0.3
0.20 B + B + 0.14
Lhole
1.8

(3.20)

andtheratiooftensiontocompressivestresses(derivedinFigure3.31)is:
103H

B =

h Y
, 0 B 10 .
Y h A h hole

(3.21)

TheplatebucklingcoefficientkBisapplicableoveralargerrangeofBthanAISIS10007
Eq. B3.25 (AISIS100 2007) and accounts for the increase in kB as the unstiffened strip
aspectratiotendstozero(i.e.,awide,shortstripresultingfromasmallhole).Referto
AppendixCforthederivationofkB.
104H

hA

fcrB

Neutral Axis

f1

Similar
Triangles

B =

f2
h Y
=
f1 Y hA hhole

Similar
Triangles

f crB
Y
=
f1 Y hA hhole

Solve for fcrB

f crhB =

f2

Y
f1
Y hA hhole

Figure3.31DerivationBandconversionofthecompressivestressattheedgeofunstiffenedstripBto
thestressfcrBattheedgeoftheplate

ConversiontothegrosssectionforthecomparisonofstressesinEq. (3.12)requires
105H

that:

forhA+hholeY,

f crh = f crh,net (1 + A )

52

hA
,
Y

(3.22)

forhA+hhole<Y,

h
h
f crh = f crh,net 1 hole 2 A hole
Y
Y

(3.23)

Theconversionfromfcrh,netatthenetsectionoftheplatetofcrhonthegrosscrosssectionis
obtainedwithasimilarmethodtothatdescribedin Figure3.14forstiffenedelementsin
106H

uniaxialcompression;thetotalcompressiveforceatthenetandgrosscrosssectionsare
assumed in equilibrium as shown in Figure 3.32 and Figure 3.33. A summary of all
107H

108H

predictionmethodequationsisprovidedinAppendixD.
109H

P
Neutral
Axis

fcrh,net

hA

fcrh
P

fcrh,net

Free Body Diagram

Force Equilibrium

+ A f crh,net
f
P = crh,net
2

Solve for fcrh

f crh = f crh ,net (1 + A )

hA
Y

1
hAt = f crhYt
2

Figure3.32DerivationoffcrhforthecasewhenhA+hholeY(whentheholeislocatedpartiallyinthe
compressedregionandpartiallyinthetensionregionoftheplate)

53

Neutral
Axis

hA

fcrh,net

f3

hhole

fcrh
P

fcrh,net

Free Body Diagram

Force Equilibrium

+ f3
f
1
1
hholet = f crhYt
P = Yf cr ,net t A crh ,net
2
2
2

f3
Y hA hhole
h
=
= A hole
f crh,net
Y
Y

Define f3 using similar


triangles

f crh = f crh,net

Substitute f3 and solve


for fcrh
Simplify

h
hhole

A f crh,net + A hole f crh,net


Y
Y

h
h
f crh = f crh,net 1 hole 2 A hole
Y
Y

Figure3.33DerivationoffcrhforthecasewhenhA+hhole<Y(holeliescompletelyinthecompressedregionof
theplate).

4.5 Verification and parameter limits


62B

The elastic buckling prediction method for stiffened elements in bending is now

evaluated with the ABAQUS eigenbuckling results presented in Section 3.4.2 and
102H

Section 3.4.3. The viability of the method is examined for evenly spaced slotted holes
102H

centered transversely or offset in a plate. Parameter limits on the prediction method,


requiredwhenformalizingthemethodforuseindesign,arealsoidentified.

ABAQUS results are compared to predictions in Figure 3.34a and Figure 3.38a.
102H

1023H

Figure3.34ademonstratesthatthesimplifiedmethodunderpredictstheelasticbuckling
1024H

stressasaspectratiooftheunstiffenedstripAincreases.Adimensionaltoleranceis
imposedtoavoidunconservativepredictions:

54

Lhole
10 .
hA

(3.24)

Eq. (3.24)alsoservesasapracticallimitontheslendernessofanunstiffenedstrip,and
1025H

thereforeisalsoisimposedontheunstiffenedstripB:

Lhole
10 .
hB

(3.25)

The prediction method becomes increasingly conservative as hA/Y approaches unity as


shownin Figure3.35a.Whenonlyasmallportionoftheholeexistsinthecompressed
1026H

regionoftheplate,theobservedbucklingmodeismoreconsistentwithplatebuckling
thanunstiffenedstripbucklingaspredictedbythesimplifiedmethod(See Figure3.28,
1027H

pictureD).Adimensionallimitissuggestedtopreventexcessiveconservatisminthis
case:

hA
0.6 *
Y

The hole spacing limits defined in Section 3.3.4.3 for stiffened elements in uniaxial

(3.26)

1028H

compressionarealsoconsideredhereforastiffenedelementinbending.Theprediction
accuracydegradeswhenholespacingSapproachestheplatewidthhasshowninFigure
1029H

3.36a. Predictions can also be unconservative when S is 2 to 3 times the length Lhole as
showninFigure3.37a.WiththelimitsfromEq.(3.8),Eq.(3.9),Eq.(3.24),Eq.(3.25),and
103H

103H

1032H

103H

1034H

Eq. (3.26)imposed,themethodisobservedtobeviablepredictoroverawiderangeof
1035H

hhole/hasshownin Figure3.38b.ThemeanandstandarddeviationoftheABAQUSto
1036H

predictedratiowithintheimposeddimensionallimitsare1.22and0.11respectively.

55

3
plate buckling
unstiffened strip "A"
unstiffened strip "B"

1.5

0.5

plate buckling
unstiffened strip "A"
unstiffened strip "B"

2.5

fcr,ABAQUS/fcr,predicted

fcr,ABAQUS/fcr,predicted

2.5

1.5

0.5

10

15

20
Lhole/hA

25

30

35

40

10

15

20
Lhole/hA

25

30

35

40

Figure3.34InfluenceofLhole/yAontheaccuracyofthepredictionmethodforstiffenedelementsinbending
(a)withoutand(b)withthedimensionallimitsdefinedinEq.(3.9),Eq.(3.24),Eq.(3.25),andEq.(3.26).

1037H

104H

plate buckling
unstiffened strip "A"
unstiffened strip "B"

2.5

fcr,ABAQUS/fcr,predicted

fcr,ABAQUS/fcr,predicted

1039H

3
plate buckling
unstiffened strip "A"
unstiffened strip "B"

2.5

1.5

0.5

1038H

1.5

0.5

0.1

0.2

0.3

0.4

0.5
hA /Y

0.6

0.7

0.8

0.9

0.1

0.2

0.3

0.4

0.5
hA /Y

0.6

0.7

0.8

0.9

Figure3.35InfluenceofhA/Yontheaccuracyofthepredictionmethodforstiffenedelementsinbending(a)
withoutand(b)withthedimensionallimitsdefinedinEq.(3.9),Eq.(3.24),Eq.(3.25),andEq.(3.26).

104H

56

1042H

1043H

104H

3
plate buckling
unstiffened strip "A"
unstiffened strip "B"

2.5

fcr,ABAQUS/fcr,predicted

fcr,ABAQUS/fcr,predicted

2.5

1.5

0.5

plate buckling
unstiffened strip "A"
unstiffened strip "B"

1.5

0.5

5
S/h

10

5
S/h

10

Figure3.36InfluenceofS/hontheaccuracyofthepredictionmethodforstiffenedelementsinbending(a)
withoutand(b)withthedimensionallimitsdefinedinEq.(3.9),Eq.(3.24),Eq.(3.25),andEq.(3.26).

1045H

1048H

plate buckling
unstiffened strip "A"
unstiffened strip "B"

2.5

fcr,ABAQUS/fcr,predicted

fcr,ABAQUS/fcr,predicted

1047H

3
plate buckling
unstiffened strip "A"
unstiffened strip "B"

2.5

1.5

0.5

1046H

1.5

0.5

5
S/Lhole

10

5
S/Lhole

10

Figure3.37InfluenceofS/Lholeontheaccuracyofthepredictionmethodforstiffenedelementsinbending(a)
withoutand(b)withthedimensionallimitsdefinedinEq.(3.9),Eq.(3.24),Eq.(3.25),andEq.(3.26).
1049H

57

105H

105H

1052H

3
plate buckling
unstiffened strip "A"
unstiffened strip "B"

1.5

0.5

plate buckling
unstiffened strip "A"
unstiffened strip "B"

2.5

fcr,ABAQUS/fcr,predicted

fcr,ABAQUS/fcr,predicted

2.5

1.5

0.5

0.1

0.2

0.3

0.4

0.5
hhole/h

0.6

0.7

0.8

0.9

0.1

0.2

0.3

0.4

0.5
hhole/h

0.6

0.7

0.8

0.9

Figure3.38Influenceofh/hholeontheaccuracyofthepredictionmethodforstiffenedelementsinbending(a)
withoutand(b)withthedimensionallimitsdefinedinEq.(3.9),Eq.(3.24),Eq.(3.25),andEq.(3.26).
1053H

1054H

105H

1056H

3.5 Unstiffenedelementinuniaxialcompression
30B

5.1 Boundary and loading conditions


63B

The unstiffened element is modeled with simplysupported boundary conditions


on three sides and unsupported on the fourth side parallel to the application of a
uniformcompressivestressasshowninFigure3.39.
1057H

Restrain plate
perimeter in 2 (v=0)
Restrain transverse
midline in 1 (u=0)

Restrain longitudinal
midline in 3 (w=0)
1

Figure3.39ABAQUSboundaryandloadingconditionsforunstiffenedplateloadeduniaxially.

58

5.2 Influence of regularly-spaced holes


64B

Eigenbuckling analyses in ABAQUS are performed to evaluate the influence of

evenlyspaced holes on the elastic buckling behavior of an unstiffened element. The


modelloadingandboundaryconditionsaresummarizedinFigure3.39andthematerial
1058H

propertiesandmeshingproceduresarethesameasthosedescribedinSection 3.2.The
1059H

platewidthh,holelengthLhole,andholetype(slotted,circular,rectangular)arevariedin
this study. The hole width remains constant at hhole=1.5 in. The plate and hole
dimensionsaswellasthecriticalelasticbucklingstress,fcrl,forthe91modelsconsidered,
areprovidedin AppendixB.Theparametricrangesconsideredinthisstudyforeach
106H

holetypearesummarizedinTable3.6.
106H

Table3.6Parameterrangeforstudyofregularlyspacedholesonunstiffenedelements.
Hole Type
Slotted
Circular
Square

hhole/h

S/Lhole

S/h

h/t

Min

0.10

1.7

1.0

21

Max

0.70

24.0

42.2

434

Min

0.10

13.3

1.3

62

Max

0.70

13.3

9.3

434

Min

0.10

13.3

1.3

62

Max

0.70

13.3

9.3

434

# of models
77
7
7

A comparison of the ABAQUS results from the 91 models to the theoretical elastic

bucklingstressforalongunstiffenedelement(k=0.425)in Figure3.40demonstratesthat
1062H

thecriticalelasticbucklingstressfcrdecreasesasholewidthhholeincreasesrelativetoplate
widthh.Holesalwaysreducethecriticalelasticbucklingstressofunstiffenedelements
in the cases studied. Buckling of the unstiffened strip A between the hole and the
simplysupportededgeisnotobservedinthesimulationsbecauseL/hisalwaysgreater
thanLhole/hA,althoughbucklingoftheunsupportedstripBatthefreeedgeoccursasthe
strip becomes slender (similar to Euler buckling) as shown in Figure 3.41. These
1063H

59

important observations are employed in Section 3.5.4 to develop an approximate


1064H

predictionmethodforthecriticalelasticbucklingstressofanunstiffenedelementwith
holes.

1.4
1.2

fcr, hole/fcr, no hole

1
0.8
0.6
0.4
0.2
0

0.2

0.4

0.6

0.8

hhole/h

Figure3.40Thepresenceofholescausesadecreaseincriticalelasticbucklingloadforunstiffenedplatesin
uniaxialcompression.

Holes do not influence the buckled shape of unstiffened plates until


the hole width becomes large relative to plate width.

hhole/h=0.10

hhole/h=0.60

hA
Buckling of the strip at the free edge of the plate changes the shape
of the local buckling mode.

Figure3.41Buckledshapesofunstiffenedplateswithholes.

60

5.3 Influence of offset slotted holes


65B

ABAQUS eigenbuckling analyses were performed to evaluate the influence of

transverselyoffsetslottedholesontheelasticbucklingofanunstiffenedelement.The
ratiooftransverseoffset, hole,toplatewidthhwasvariedfrom0.375to0.375,wherea
negativeoffsetshiftstheholestowardthesimplysupportededgeandapositiveoffset
shiftstowardsthefreeplateedge(referto Figure3.2foradefinitionof hole).Themodel
1065H

loading and boundary conditions are summarized in Figure 3.39 and the material
106H

propertiesandmeshingproceduresarethesameasthosedescribedinSection 3.2.The
1067H

plate and hole dimensions as well as the critical elastic buckling stress, fcrl, for the 92
modelsconsidered,aresummarizedin AppendixB.Theparametricrangesconsidered
1068H

hereareprovidedinTable3.7.
1069H

Table3.7Parameterrangeconsideredforunstiffenedelementstudywithoffsetholes
Slotted

hhole/h

S/Lhole

S/h

h/t

hole/h

Min

0.10

5.00

1.33

62

-0.375

Max

0.70

5.00

9.33

434

0.375

# of models
92

The axial stiffness of an unstiffened element is higher near the simply supported

edge and lower near the free edge. It is hypothesized that holes shifted towards the
simplysupported edge will reduce the critical elastic buckling stress more than hole
materialremovedfromnearthefreeedge.Thishypothesisisconfirmedin Figure3.43
107H

where fcr decreases more when holes are shifted towards the simplysupported edge.
The dimension of the plate strip between the hole and the simplysupported edge, hA
(see Figure3.3),isidentifiedasausefulparameterwhenpredictingfcr.fcrformsatrend
107H

line when plotted against Lhole relative to yA as demonstrated in Figure 3.43a for offset
1072H

61

holes.ThesameplotisproducedusingthedatafromSection 3.5.2forcenteredholesin
1073H

Figure 3.43b with similar results. This important conclusion, that yA and Lhole are key
1074H

parametersinfluencingfcr,isusedinthenextsectiontodevelopanapproximateelastic
bucklingpredictionmethodforunstiffenedelementswithholes.

1.4
1.2

fcr, hole/fcr, no hole

1
0.8
Data groups
correspond to
hhole/h=0.10, 0.20,
0.70

0.6
0.4
0.2
0
-0.4

-0.3

-0.2

-0.1

0.1

0.2

0.3

0.4

hole/h

Figure3.42Thecriticalelasticbucklingstressofastiffenedplatedecreasesasholesareshiftedtowardthe

1.4

1.4

1.2

1.2

1
fcr, hole/fcr, no hole

fcr, hole/fcr, no hole

simplysupportededge(+hole)

0.8
0.6

0.8
0.6

0.4

0.4

0.2

0.2

10

15

20
Lhole/hA

25

30

35

40

10

15

20
Lhole/hA

25

30

35

40

Figure3.43Thecriticalelasticbucklingstressforstiffenedelementswith(a)transverselyoffsetholesand
(b)centeredholes(fromSection3.5.2)decreasesasafunctionofholelengthLholetohA

1075H

62

5.4 Approximate prediction method for use in design


6B

Anapproximateelasticbucklingpredictionmethodforanunstiffenedelementwith

holes is presented here. The method is based on the observations in Section 3.3.2 and
1076H

Section 3.3.3forlongunstiffenedelementswithevenlyspacedholes.Thewidthofthe
107H

stripbetweentheholeandthesimplysupportededge,hA,andthelengthoftheholeLhole
are utilized as predictors of the critical elastic buckling stress. A summary of the
predictionmethodequationsareprovidedinAppendixD.
1078H

5.4.1 Derivationofempiricalbucklingcoefficient
123B

An empirical plate buckling coefficient is determined using a linear regression

analysisofthedatain Figure3.43aand Figure3.43bforbothcenteredandoffsetslotted


1079H

108H

holes, which was then adjusted to have a slightly conservative bias. The regression
minimizestheerrorbetweentheABAQUSresultsandtheclassicalstabilitysolutionof
an unstiffened element (k=0.425) for the plate models within the following parametric
limits:

Lhole
10
hA

(3.27)

Lhole
10
hB

(3.28)

hhole
0.50
h

(3.29)

Eq. (3.27) is imposed as a practical limit on the slenderness of the strip adjacentto the
108H

hole at the simplysupported plate edge. Eq. (3.28) prevents Euler buckling of
1082H

63

unstiffened strip B as shown in Figure 3.41. Eq. (3.29) is imposed because of the
1083H

1084H

increasedrateofdegradationinfcrobservedinFigure3.40asholesbecomelargerelative
1085H

toplatewidth.Theempiricalplatebucklingcoefficientissetas:

L
k = 0.4251 0.062 hole
hA

(3.30)

where the strip of plate between the hole and the simply supported edge, hA, is
calculatedas

hA =

h hhole
hole .
2

(3.31)

Apositivehole(holeoffsetfromthecenterlineoftheplate,SeeFigure3.2)shiftsthehole
1086H

towardsthesimplysupportededge.Theempirical bucklingcoefficientinEq. (3.30)is


1087H

shown in Figure 3.44a to be a slightly conservative but accurate representation of


108H

ABAQUS predicted buckling coefficients. The mean and standard deviation of the
ABAQUStoempiricalpredictionratioare1.06and0.09respectively.
0.5

1.4
ABAQUS
Eq. (3.30)

0.45

1.2

0.4
1
fcr, ABAQUS/fcr, predicted

buckling coeff. k

0.35
0.3
0.25
0.2
0.15

0.8

0.6

0.4

0.1
0.2

0.05
0

5
Lhole/y A

10

5
Lhole/y A

10

Figure3.44(a)ComparisonofABAQUSandempiricalplatebucklingcoefficientsforanunstiffenedelement
withholesand(b)ABAQUStopredictedelasticbucklingstressforanunstiffenedelement

64

5.4.2 Predictionequations
124B

The elastic buckling stress of an unstiffened element in compression with holes is

thusapproximatedas:

f crl = min[ f cr , f crh ] .

(3.32)

The critical elastic buckling stress prediction for plate buckling of the unstiffened
elementwithoutholes(fcr)iscalculatedwithEq. (3.2),wherek=0.425whenconsidering
1089H

long rectangular plates (L/h>4). The minimum critical elastic buckling stress of the
unstiffened element with holes, fcrh, coincides with either buckling of the entire
unstiffenedelementwithholesorbucklingoftheunstiffenedstripAadjacenttothe
holeandthesimplysupportededge,or:

f crh

2E t
h
, f crA 1 hole
= min k
2
h

12 1 h

(3.33)

where k is an empirical plate buckling coefficient derived from finite element


eigenbucklingstudiesinEq. (3.30).fcrAiscalculatedwithEq. (3.4)andmodifiedbythe
109H

109H

factor(1hhole/h)toconvertthestressontheunstiffenedstripAtothestressattheend
oftheplatesothatitcanbecomparedtothebucklingstressoftheunstiffenedelement.
fcrhwillalwaysbepredictedaslessthanorequaltofcrwiththismethod.

65

Chapter 4
Elastic buckling of cold-formed steel
members with holes
3B

The elastic buckling properties of coldformed steel lipped Csection beams and

columns with holes are evaluated in this chapter using thinshell finite element
eigenbucklinganalysesinABAQUS.Theelasticbucklingstudiesareusedtoassessthe
influenceofholesonthelocal,distortional,andglobalcriticalelasticbucklingloadsPcrl,
Pcrd, Pcre. The studies also identify elastic buckling modes unique to coldformed steel
memberswithholes.Elasticbucklingpropertiesofexistingexperimentsoncoldformed
steel columns and beams with holes are summarized and formal buckling modes are
definedinpreparationforthepresentationoftheDirectStrengthMethodforstructural
memberswithholesinChapter8.
1092H

66

4.1 Finiteelementmodelingassumptions
31B

Theelasticbucklingbehaviorofthecoldformedsteelstructuralmemberswithholes

areobtainedwitheigenbucklinganalysesinABAQUS(ABAQUS2007a).Allmembers
are modeled with ABAQUS S9R5 reduced integration ninenode thin shell elements.
Thetypicalfiniteelementaspectratiois1:1andthemaximumaspectratioislimitedto
8:1(referto Chapter2foradiscussiononABAQUSthinshellfiniteelementtypesand
1093H

finite element aspect ratio limits). Element meshing is performed with a Matlab
(Mathworks2007)programwrittenbytheauthor(refertoAppendixAforadescription
1094H

oftheprogram).ColdformedsteelmaterialpropertiesareassumedasE=29500ksiand
=0.3 in the finite element models unless noted otherwise. Py, the squash load of the
column,iscalculatedbymultiplyinganassumedyieldstressof50ksibythegrosscross
sectionalareaofthecolumn.

4.2 Elasticbucklingofcolumnswithholes
32B

2.1 Member and hole dimensions


67B

Memberandholedimensionnotationusedthroughoutthischapteris summarized

in Figure4.1.Uppercasedimensions(H,D,B)areouttooutandlowercasedimensions
1095H

(b,h)areflatlengthsbetweenpointsofcurvature.

67

B1
D1

bhole

hhole

t
r

D2
b

B2

Figure4.1Csectionandholedimensionnotation

2.2 Loading and boundary conditions


68B

Thecoldformedsteelcolumnboundaryconditionsaremodeledaswarpingfreeat

thememberendsandwarpingfixedatthemidlengthofthememberasshowninFigure
1096H

4.2,whichmimicsthesemianalyticalfinitestripmethod(Schaferanddny2006).The
columns are loaded at each end with stress distributions applied as consistent nodal
loadsinABAQUS(seeSection3.2fordetailsontheloadingimplementation).
1097H

68

midspan nodes are


supported in 1 (u = 0)
end nodes are
supported in 2 and
3 (v = w = 0)

end nodes are supported


in 2 and 3 (v = w = 0)

Figure4.2Columnsaremodeledwithpinnedwarpingfreeboundaryconditionsandcompressedfrom
bothends

2.3 Elastic buckling comparison of short C-section columns versus


69B

isolated stiffened elements

This study builds on the results and observations in Chapter 3 for crosssectional
1098H

elementswithholesandmarksatransitioninresearchfocusfromelementstofullcold
formed steel members. The influence of one slotted hole on the elastic buckling
behavior of a range of rectangular plates and SSMA coldformed steel structural stud
sectionsiscompared,thegoalbeingtoquantifytherelativeinfluenceofawebholeon
one element in a crosssection (in this case a stiffened element, see Figure 3.1 for
109H

definitionandFigure3.5forABAQUSboundaryconditions)versusafullCsection.The
10H

slottedholehasdimensionsofhhole=1.5in.,Lhole=4in.,andrhole=0.75in.Theplatewidths
are chosen to correspond with the flat web widths of standard SSMA structural studs
(SSMA 2001). Plate aspect ratios are held constant at 4:1. From each plate, a full
structural stud finite element model is developed for comparison. The SSMA member

69

designations and cross section dimensions considered in this study are listed in Table
10H

4.1.
Table4.1SSMAstructuralstudandplatedimensions

SSMA
Designation
250S162-33
350S162-33
362S162-33
400S162-33
550S162-33
600S162-33
800S162-33

H
in.
2.50
3.50
3.62
4.00
5.50
6.00
8.00

B
in.
1.63

D
r
t
in.
in.
in.
0.50 0.0764 0.0346

b hhole/h
in.
1.40 0.66
0.46
0.44
0.40
0.28
0.26
0.19

h
in.
2.28
3.28
3.40
3.78
5.28
5.78
7.78

L=4h
in.
9.1
13.1
13.6
15.1
21.1
23.1
31.1

Beforeexaminingtheelasticbucklingload,considertheobservedchangesinthefirst

mode shape caused by the addition of a hole as given in Figure 4.3. For the buckled
102H

shapeoftheSSMA250S16233inFigure4.3a,thenumberofbuckledhalfwaveschanges
103H

fromfourtothreefortheisolatedplateandfromfivetotwoforthefullmember,when
theholeisadded.Thestripsofplateadjacenttotheholearestiffenedbytheconnected
flangeinthefullmember,causingbuckledhalfwavestoforminthewebawayfromthe
hole.Also,thelengthofthehole,Lhole,isapproximatelyhalfofthelengthofthemember
L in the SSMA 250S16233 member which also prevents local buckling in the web. In
Figure4.3b,theholedecreasesthenumberofbuckledhalfwavesfromfourtothreein
104H

theSSMA440S16233isolatedplatebutdoesnotchangethenumberofhalfwavesinthe
fullmember.Thecrosssectionconnectivityofthefullmemberlimitsdeformationatthe
holeandencouragesbucklinghalfwavestoformalongtheentiremember.Also,there
ismorewebmaterialtoaccommodatelocalbucklingalongthelength(Lhole/L=0.26)when
comparedtotheSSMA250S16233member.

70

# of Local Buckling
Half-Waves (Typ.)
4

3
2

3
5

(a)

(b)
SSMA 400S162-33
hhole/h=0.40

SSMA 250S162-33
hhole/h=0.66

Figure4.3(a)SSMA250S16233webplateandstructuralstud,and(b)SSMA400S16233webplateand
structuralstud

Figure 4.4 presents the influence of a slotted hole on the critical elastic buckling
105H

stress fcr of the isolated web plates and full members with holes from Table 4.1. These
106H

resultsarecomparedtotheelasticbucklingpredictionforastiffenedelementwithholes
developedandpresentedinSection 3.3.4.Theinfluenceoftheholeisminimalforsmall
107H

hole width to plate width ratios, but increases to a maximum at hhole/h=0.30 for the
ABAQUS plate results (consistent with the stiffened element prediction). fcr increases
withincreasinghhole/hforfullmembers,demonstratingthatthecrosssectionconnectivity
decreases a members sensitivity to a hole (especially in the range of hhole/h=0.30). The
webisstiffenedthroughbeneficialwebflangeinteractioncreatedbytherelativelystable
edgestiffenedflange.

Asnormalizedholewidthincreases,theelasticbucklingloadexceedsthatofaplate

withoutahole.Thisincreaseinbucklingloadisattributedtotheincreasedaxialstiffness
fromthestripsadjacenttothehole,whichcauseslocalbucklingtooccurawayfromthe
holes (this unstiffened strip effect is discussed in Section 3.3). The length of hole
108H

relativetothelengthofthememberalsocontributestotheincreaseinfcr.Lhole/Lincreases

71

with increasing hhole/h in this study since L=4h. As demonstrated in Figure 4.3a, the
109H

removal of web material restricts the formation of local buckling in the web of the
member, resulting in shortened halfwaves away from the hole with increased axial
stiffness. The stiffened element prediction in Figure 4.4 is conservative here when
10H

comparedtotheplateresultsbecauseitwasdevelopedforevenlyspacedholesinlong
platesanddoesnotaccountfortheLhole/Lboostinfcr.

1.6
ABAQUS, SSMA Stud
ABAQUS, stiffened element
Stiffened element prediction

1.4

fcr,hole/fcr,no hole

1.2
1
0.8
0.6
0.4
0.2
0

0.1

0.2

0.3

0.4

0.5
hhole/h

0.6

0.7

0.8

0.9

Figure4.4.Effectofaslottedholeontheelasticbucklingloadofsimplysupportedplatesandstructural
studs

2.4 Influence of slotted hole location on elastic buckling of an


70B

intermediate length structural stud

Thisstudyinvestigatestheelasticbucklingbehaviorofanintermediatelengthcold

formed steel column with one slotted hole. The primary goal here is to identify and
formally define the elastic buckling modes that will be used as predictors of ultimate
strengthwithintheDirectStrengthMethod.Theelasticbucklingbehavioriscompared

72

as the location of a slotted hole is varied along the length of the column. The typical
compressionmemberinthisstudyhasalength L of48inchesandismodeledwithan
SSMA 36216233 structural channel cross section. A single slotted hole is centered
transverselyintheweb.Theslottedholehasdimensionsofhhole=1.5in.,Lhole=4in.,and
rhole=0.75in. Table4.1summarizesthedimensionsoftheSSMA36216233crosssection.
1H

TheABAQUScolumnboundaryconditionsareconsistentwithFigure4.2.
12H

Figure 4.5 compares the local buckling (L) mode shapes of the column with and
13H

without a slotted hole. The lowest buckling mode is local buckling (L) and exists for
boththecolumnwithandwithoutthehole.Thelocationoftheholedoesnotinfluence
PcrforthismodeasobservedinFigure4.8.Platebucklingofthewebawayfromthehole
14H

dominatesforthismode,regardlessofholelocation.
Two unique local buckling modes to the column with a hole, LH and LH2, are
alsoidentifiedintheeigenbucklinganalyses.Thesemodes,showninFigure4.5,exhibit
15H

buckling of the unstiffened strip adjacent to the hole similar to that observed in the
crosssectionalelementstudieswithholesinChapter3.TheLHmodeoccurswhenboth
16H

unstiffenedstripsbuckleinthesamedirectionnormaltothewebplane,whichincreases
the distortional tendencies of the flange in the vicinity of the hole. This localized
distortional buckling is observed in Figure 4.6, which compares the LH modes as the
17H

locationoftheholevariesalongthecolumnlength.TheLHmodeisconsistentwiththe
elastic buckling mode shapes of stiffened elements, where the presence of a hole is
observedtoreducethetransversebendingstiffnesscausinglocalizeddeformationatthe
hole(seeFigure3.9a).
18H

73

TheLH2modeoccurswhentheunstiffenedstripsbuckleinoppositedirections
relativetothewebplate,resultinginantisymmetricdistortionaldeformationatthehole.
Figure4.8demonstratesthatPcrforthesetwomodesissimilarandthatbothmodesare
19H

minimallyaffectedbythelongitudinallocationoftheholeinthecolumn.

LH

LH2

Figure4.5Thepresenceofaholecreatesuniquelocalbucklingmodeswhereunstiffenedstripbuckling
adjacenttotheholeoccurssymmetrically(LH)orasymmetrically(LH2)increasethedistortionaltendency
oftheflanges

C
L hole

Figure4.6SSMAslottedholelocationandlocalbucklingLHmode,L=48in.,x/L=0.06,0.125,0.25,0.375,0.50.
Notethedistortionaltendenciesoftheflangesatthehole.

74

The pure distortional buckling mode for the column, D, is compared for a column

with and without a hole in Figure 4.7a. Note that distortional halfwavelength is
120H

unchangedwiththepresenceofthehole,althoughlocalbucklingwithhalfwavelengths
shorter than the fundamental L halfwavelength (see Figure 4.5) mix with the D mode
12H

forthememberwiththehole.Figure4.8demonstratesthatthepresenceoftheholehas
12H

aminimalinfluenceonPcrdregardlessoflongitudinallocation.

The lowest global buckling mode is flexuraltorsional buckling (GFT) as shown in

Figure4.7b.ThepresenceofaholeresultsinaslightdecreaseinPcreastheholemoves
123H

towards the end of the column as shown in Figure 4.8. As the hole shifts close to the
124H

loadedendofthecolumn(x/L=0.06),localbucklingattheholecombineswiththeGFT
modetoreducePcre.

GFT

Figure4.7Influenceofaslottedholeonthe(a)distortional(D)and(b)globalflexuraltorsional(GFT)modes
ofacoldformedsteelcolumn

75

1
D
GFT
LH2
LH
L

0.9
0.8

D, no hole

0.7

Pcr/Py

0.6

GFT, no hole

0.5
0.4
0.3
0.2

L, no hole

0.1
0

0.1

0.2

0.3

0.4
x/L

0.5

0.6

0.7

0.8

Figure4.8InfluenceofSSMAslottedholelocationonPcrfora362S16233Csection(refertoFigure4.5,
Figure4.6,andFigure4.7forbuckledshapesummaries)
125H

126H

127H

2.5 Flange hole study


71B

The research focus up until now has been on the elastic buckling modes of
isolatedwebplatesandcoldformedsteelcompressionmemberswithwebholes.Holes
arealsocommonlypresentintheflangesofCsectioncolumns.Astandardconnection
detail requiring a flange hole is shown in Figure 4.9, where gypsum sheathing is
128H

connected to steel studs with a bolted or screw attachment (Western States Clay
ProductsAssociation2004).

76


Figure4.9Connectiondetailforstructuralstudtoexteriorwallrequiresascreworboltholeplacedinthe
studflange(WesternStatesClayProductsAssociation2004)

Thisstudyevaluatestheinfluenceofcircularflangeholesontheelasticbuckling
behavior of an intermediate length SSMA 362S16233 structural stud. A single hole is
placed at the midlength of both the top and bottom flanges and centered between the
web and lip stiffeners. Five hole diameters consistent with standard bolt holes are
considered:bhole/b=0.178,0.356,0.534,0.713,and0.892(,,,1,1holesina1
flange) where the flat flange width b=1.40 in. The length L is 48 in. for all members,
correspondingtoacommonunbracedlengthofaSSMAstructuralstud.
Figure 4.10 presents the variation in elastic buckling loads for local, distortional,
129H

and global modes as the diameter of the flange holes increase. The local (L) buckling
load, Pcrl, is not influenced by small holes, but decreases as bhole/b exceeds 0.70. Figure
130H

4.11 demonstrates that for large flange holes local buckling is dominated by web and
flangedeformationneartheholes.Thelargeflangeholesadverselyaffectthebeneficial
webflangeinteractioninherentinstructuralstuds(Figure 4.4highlightsthisbeneficial
13H

interaction for Csections with web holes). The interruption of the webflange
interaction by the holes is also reflected in the pure distortional mode (D), as Pcrd

77

decreases slightly as flange hole size increases relative to flange width. The flanges
holes have a minimal effect on the global flexuraltorsional mode (GFT) because their
diameterissmallrelativetothecolumnlength.

1
D
GFT
L

0.9
0.8

D, no hole

0.7

Pcr/Py

0.6

GFT, no hole

0.5
0.4

L, no hole

0.3
0.2
0.1
0

0.1

0.2

0.3

0.4

0.5
bhole/b

0.6

0.7

0.8

0.9

Figure4.10Influenceofflangeholediameteronthelocal(L),distortional(D),andglobal(GFT)elastic
bucklingloadsofanSSMA362S16233structuralstud

Figure4.11Local(L)bucklingisdominatedbyflangeandwebdeformationneartheholesasbhole/bexceeds
0.70

78

4.2.6 Analysis of existing experiments on columns


72B

TheDirectStrengthMethodemploystheelasticbucklingpropertiesofacoldformed

steel member to predict its ultimate strength. To assist in the extension of DSM to
columnswithholes,adatabaseisdevelopedinthissectionwhichsummarizestheelastic
bucklingpropertiesandtestedstrengthsofcoldformedsteelcolumnsexperimentswith
holes from the past 30 years. This database is used Chapter 8 when developing and
132H

verifying DSM for columns with holes. Table 4.2 summarizes the experimental
13H

programscomprisingthedatabase.
Table4.2Summaryofcolumnexperimentaldata

Author
Ortiz-Colberg
Ortiz-Colberg
Miller and Pekz
Sivakumaran
Abdel-Rahman
Pu et al.
Moen and Schafer
Moen and Schafer

Publication Date
1981
1981
1994
1987
1997
1999
2008
2008

Types of Specimens
Stub Column
Long Column
Stub Column
Stub Column
Stub Column
Stub Column
Short Column
Intermediate Column

Cross Section
Lipped Cee Channel

End Conditions
Fixed-Fixed
Weak axis pinned
Fixed-Fixed
Fixed-Fixed
Fixed-Fixed
Fixed-Fixed
Fixed-Fixed
Fixed-Fixed

# of Specimens
8
15
12
14
8
9
6
6

2.6.1 Elasticbucklingdatabaseforcolumnexperiments
125B

ABAQUSeigenbucklinganalyseswereconductedforeachspecimeninthedatabase.

Member boundary conditions and loading conditions were modeled to be consistent


with the actual experimental conditions. The ABAQUS implementations of the
boundaryconditionsforeachexperimentalprogramaredescribedin Figure4.12.Local
134H

(L, LH, LH2 see Figure 4.5), distortional (D), and global (G) buckling modes were
135H

manually identified from the buckled modes in ABAQUS using the modal definitions
describedin4.2.4.
136H

79

Nodes bearing on top platen


constrained in 1, 2 and 3

Boundary conditions valid for:

Boundary conditions valid for:


Boundary conditions valid for:

Ortiz-Colberg 1981 (stub columns)


Sivakumaran 1987
Pu et al. 1999
Moen and Schafer 2008

Miller and Pekoz 1994


Abdel-Rahman 1997

Specimens bear directly on the


platens, ends are not welded

1
2

ABAQUS pinned rigid body


reference node constrained in 2 to 6
directions, ensures that all nodes on
loaded surface translate together in 1
direction but can rotate freely at the
platen (no welding)

4
6

ABAQUS tied rigid body reference


node constrained in 1, 3 to 6 directions,
ensures that all nodes on contact
surface translate together in 1 direction
and cannot rotate at the platen (crosssection contact edge is welded). The
entire platen can rotate in the 2
direction.

Nodes bearing on top platen


constrained in 1 to 6
(contact edge is welded)

Ortiz-Colberg 1981 (long column)

Specimens are welded to loading


platens, cross-section edge rotation is
restrained at contact location

Specimens are welded to loading


platens, which are attached to pins that
allow weak axis rotation of the platens

ABAQUS tied rigid body reference


node constrained in 2 to 6 directions,
ensures that all nodes on loaded
surface translate together in 1 direction
and cannot rotate at the platen (crosssection contact edge is welded)

ABAQUS tied rigid body reference node


constrained in 3 to 6 directions, ensures that
all nodes on loaded surface translate
together in 1 direction and cannot rotate at
the platen (cross-section contact edge is
welded). The entire platen can also rotate in
the 5 direction.

(a)

(b)

(c)

Figure4.12ExperimentalprogramboundaryconditionsasimplementedinABAQUS

Table 4.3 summarizes the fixedfixed column specimen dimensions and material
137H

properties,includingcrosssectionandholedimensions,testedultimateloadPtest,tested
specimen yield stress Fy, specimen yield force Py,g (calculated with the gross cross
sectional area), and Py,net (calculated with the net crosssectional area). Table 4.4
138H

summarizes the fixedfixed column specimen elastic buckling loads. ABAQUS


eigenbucklingresultsarepresentedfortwodifferenttypesofboundaryconditions,the
experimentboundaryconditionsandCUFSMboundaryconditions(warpingfreeatthe
ends, warpingfixed at the midlength of the column) except for the Moen and Schafer
specimens which were only modeled with experiment boundary conditions. CUFSM
elasticbucklingresultsarealsoprovided,includingthedistortionalhalfwavelengthLcrd.
Thesameexperimentandelasticbucklinginformationissummarizedfortheweakaxis
pinned columns in Table 4.5 and Table 4.6 and together with Table 4.3 and Table 4.4
139H

140H

14H

142H

comprisethedatasetusedtoevaluatetheDSMequationsforcoldformedsteelcolumns
withholesChapter8.
143H

80

Table4.7summarizescrosssectionandmaterialpropertyparameterrangesforthe
14H

fixedfixedspecimensandweakaxispinnedspecimens.Mostoftheweakaxispinned
specimens are long columns, while the majority of the fixedfixed specimens are stub
columns (the exception being the short and intermediate length fixedfixed columns
tested by Moen and Schafer). All column specimens in the database are common
industry shapes and meet the DSM prequalification standards (for members without
holes)summarizedinTable4.8(AISIS1002007).
145H

81

Table4.3Fixedfixedcolumnexperimentdimensionsandmaterialproperties

Member
Study and Specimen Name
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Abdel-Rahman 1997
Abdel-Rahman 1997
Abdel-Rahman 1997
Abdel-Rahman 1997
Abdel-Rahman 1997
Abdel-Rahman 1997
Abdel-Rahman 1997
Abdel-Rahman 1997
Pu et al. 1999
Pu et al. 1999
Pu et al. 1999
Pu et al. 1999
Pu et al. 1999
Pu et al. 1999
Pu et al. 1999
Pu et al. 1999
Pu et al. 1999
Sivakumaran 1987
Sivakumaran 1987
Sivakumaran 1987
Sivakumaran 1987
Sivakumaran 1987
Sivakumaran 1987
Sivakumaran 1987
Sivakumaran 1987
Sivakumaran 1987
Sivakumaran 1987
Sivakumaran 1987
Sivakumaran 1987
Sivakumaran 1987
Sivakumaran 1987
Miller & Pekoz 1994
Miller & Pekoz 1994
Miller & Pekoz 1994
Miller & Pekoz 1994
Miller & Pekoz 1994
Miller & Pekoz 1994
Miller & Pekoz 1994
Miller & Pekoz 1994
Miller & Pekoz 1994
Miller & Pekoz 1994
Miller & Pekoz 1994
Miller & Pekoz 1994
Moen and Schafer 2008
Moen and Schafer 2008
Moen and Schafer 2008
Moen and Schafer 2008
Moen and Schafer 2008
Moen and Schafer 2008
Moen and Schafer 2008
Moen and Schafer 2008
Moen and Schafer 2008
Moen and Schafer 2008
Moen and Schafer 2008
Moen and Schafer 2008
* average of two tests

S4
S7
S6
S8
S5
S3
S14
S15
A-C
A-S
A-O
A-R
B-C
B-S
B-O
B-R
C-2.0-1-30-1
C-2.0-1-30-2
C-2.0-1-30-3
C-1.2-1-30-1
C-1.2-1-30-2
C-1.2-1-30-3
C-0.8-1-30-1
C-0.8-1-30-2
C-0.8-1-30-3
A2
A3
A4
A5
A6
A7
A8
B2
B3
B4
B5
B6
B7
B8
1-12
1-13
1-17
1-19
2-11
2-12
2-14
2-15
2-16
2-24
2-25
2-26
362-1-24-H
362-2-24-H
362-3-24-H
362-1-48-H
362-2-48-H
362-3-48-H
600-1-24-H
600-2-24-H
600-3-24-H
600-1-48-H
600-2-48-H
600-3-48-H

Boundary
Conditions

in.
Fixed-fixed 12.00
12.00
12.00
12.00
12.00
12.00
12.00
12.00
16.73
16.73
18.70
18.70
9.84
9.84
11.81
11.81
14.57
14.57
14.57
14.17
14.17
14.17
14.17
14.17
14.17
7.87
7.87
7.87
7.87
7.87
7.87
8.78
10.43
10.43
10.43
10.43
10.43
10.43
10.43
10.87
10.87
17.95
17.95
10.87
10.87
17.95
17.95
17.95
17.95
17.95
17.95
24.1
24.1
24.1
48.22
48.23
48.2
24.1
24.1
24.1
48.09
48.25
48.06

Material

Hole Dimensions

Cross Section Dimensions

Yield Stress and Force

Test

nu

Hole Type

L hole

h hole

r hole

B1

B2

D1

D2

Fy

P y,g

P y,net

P test

in.
0.0492
0.0493
0.0496
0.0496
0.0498
0.0499
0.0760
0.0760
0.0740
0.0740
0.0740
0.0740
0.0500
0.0500
0.0500
0.0500
0.0787
0.0787
0.0787
0.0472
0.0472
0.0472
0.0315
0.0315
0.0315
0.0630
0.0630
0.0630
0.0630
0.0630
0.0630
0.0630
0.0508
0.0508
0.0508
0.0508
0.0508
0.0508
0.0508
0.0756
0.0752
0.0346
0.0346
0.0752
0.0752
0.0350
0.0346
0.0350
0.0354
0.0354
0.0350
0.0391
0.0383
0.0394
0.0393
0.0391
0.0399
0.0421
0.0412
0.043
0.0428
0.0429
0.0431

ksi
29420
29420
29420
29420
29420
29420
29420
29420
29420
29420
29420
29420
29420
29420
29420
29420
29420
29420
29420
29420
29420
29420
29420
29420
29420
29710
29710
29710
29710
29710
29710
29710
30435
30435
30435
30435
30435
30435
30435
29420
29420
29420
29420
29420
29420
29420
29420
29420
29420
29420
29420
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500

0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3

Circular
Circular
Circular
Circular
Circular
Circular
Circular
Circular
Circular
Square
Oval
Rectangle
Circular
Square
Oval
Rectangle
Square
Square
Square
Square
Square
Square
Square
Square
Square
Circular
Square
Circular
Square
Circular
Square
Oval
Circular
Square
Circular
Square
Circular
Square
Oval
Rectangular
Rectangular
Rectangular
Rectangular
Rectangular
Rectangular
Rectangular
Rectangular
Rectangular
Rectangular
Rectangular
Rectangular
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted

in.
0.75
1.50
1.25
1.75
1.04
0.50
1.04
1.50
2.50
2.50
4.50
4.50
1.50
1.50
4.00
4.00
1.06
1.05
1.05
1.04
1.04
1.04
1.04
1.04
1.04
0.65
0.65
1.30
1.30
1.95
1.95
4.02
1.14
1.14
2.28
2.28
3.43
3.43
4.02
2.76
2.76
2.24
2.24
2.56
2.56
2.24
2.24
2.24
2.24
2.24
2.24
4.00
4.00
4.01
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00

in.
0.75
1.50
1.25
1.75
1.04
0.50
1.04
1.50
2.50
2.50
2.50
2.50
1.50
1.50
1.50
1.50
1.06
1.05
1.05
1.04
1.04
1.04
1.04
1.04
1.04
0.65
0.65
1.30
1.30
1.95
1.95
1.50
1.14
1.14
2.28
2.28
3.43
3.43
1.50
1.61
1.61
1.57
1.57
1.50
1.50
1.57
1.57
1.57
1.57
1.57
1.57
1.49
1.50
1.49
1.50
1.50
1.49
1.50
1.49
1.49
1.49
1.50
1.50

in.
0.38
0.75
0.63
0.88
0.52
0.25
0.52
0.75
1.25
--1.25
--0.75
--0.75
--------------------0.32
--0.65
--0.97
--0.75
0.57
--1.14
--1.71
--0.75
------------------------0.75
0.75
0.75
0.75
0.75
0.75
0.75
0.75
0.75
0.75
0.75
0.75

in.
3.50
3.51
3.51
3.51
3.50
3.50
3.52
3.52
7.99
7.99
7.99
7.99
4.00
4.00
4.00
4.00
3.94
3.94
3.94
3.87
3.87
3.87
3.84
3.84
3.84
3.63
3.63
3.63
3.63
3.63
3.63
3.63
6.00
6.00
6.00
6.00
6.00
6.00
6.00
3.62
3.62
5.98
5.98
3.62
3.62
5.98
5.98
5.98
5.98
5.98
5.98
3.58
3.64
3.67
3.62
3.62
3.63
6.04
6.01
6.03
6.01
6.02
6.06

in.
1.62
1.63
1.61
1.62
1.62
1.61
1.67
1.67
1.64
1.64
1.64
1.64
1.64
1.64
1.64
1.64
2.05
2.05
2.05
2.05
2.05
2.05
2.05
2.05
2.05
1.63
1.63
1.63
1.63
1.63
1.63
1.63
1.63
1.63
1.63
1.63
1.63
1.63
1.63
1.46
1.46
1.34
1.34
1.46
1.46
1.38
1.38
1.38
1.38
1.38
1.38
1.65
1.63
1.67
1.60
1.59
1.60
1.59
1.61
1.61
1.60
1.59
1.63

in.
1.49
1.49
1.48
1.48
1.48
1.48
1.49
1.49
1.64
1.64
1.64
1.64
1.64
1.64
1.64
1.64
2.05
2.05
2.05
2.05
2.05
2.05
2.05
2.05
2.05
1.63
1.63
1.63
1.63
1.63
1.63
1.63
1.63
1.63
1.63
1.63
1.63
1.63
1.63
1.46
1.46
1.34
1.34
1.46
1.46
1.38
1.38
1.38
1.38
1.38
1.38
1.60
1.59
1.70
1.60
1.61
1.61
1.61
1.60
1.58
1.63
1.61
1.59

in.
0.49
0.50
0.49
0.49
0.49
0.48
0.51
0.51
0.51
0.51
0.51
0.51
0.51
0.51
0.51
0.51
0.63
0.63
0.63
0.63
0.63
0.63
0.63
0.63
0.63
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.47
0.47
0.31
0.31
0.47
0.47
0.31
0.31
0.31
0.31
0.31
0.31
0.43
0.44
0.42
0.42
0.42
0.40
0.48
0.37
0.36
0.48
0.48
0.37

in.
0.50
0.51
0.51
0.50
0.50
0.50
0.51
0.51
0.51
0.51
0.51
0.51
0.51
0.51
0.51
0.51
0.63
0.63
0.63
0.63
0.63
0.63
0.63
0.63
0.63
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.47
0.47
0.31
0.31
0.47
0.47
0.31
0.31
0.31
0.31
0.31
0.31
0.44
0.39
0.43
0.41
0.40
0.43
0.36
0.50
0.48
0.39
0.36
0.48

in.
0.10
0.10
0.10
0.10
0.10
0.10
0.10
0.10
0.15
0.15
0.15
0.15
0.10
0.10
0.10
0.10
0.16
0.16
0.16
0.11
0.11
0.11
0.08
0.08
0.08
0.13
0.13
0.13
0.13
0.13
0.13
0.13
0.10
0.10
0.10
0.10
0.10
0.10
0.10
0.09
0.09
0.09
0.09
0.09
0.09
0.09
0.09
0.09
0.09
0.09
0.09
0.20
0.20
0.19
0.19
0.19
0.18
0.16
0.15
0.16
0.16
0.16
0.16

ksi
47.1
48.5
51.5
51.5
49.6
49.6
47.4
47.6
55.8
55.8
55.8
55.8
46.2
46.2
46.2
46.2
44.4
33.6
34.4
28.0
28.0
28.0
24.8
24.8
24.8
49.4
49.4
49.4
49.4
49.4
49.4
49.4
38.1
38.1
38.1
38.1
38.1
38.1
38.1
51.9
51.9
44.8
44.8
53.0
53.0
43.8
43.8
43.8
43.8
43.8
43.8
57.9
57.1
56.0
58.6
59.7
58.3
61.9
58.4
60.1
61.4
62.0
61.5

kips
16.7
17.3
18.4
18.4
17.7
17.7
25.8
25.9
48.3
48.3
48.3
48.3
18.2
18.2
18.2
18.2
30.2
22.8
23.4
11.6
11.6
11.6
7.0
7.0
7.0
22.9
22.9
22.9
22.9
22.9
22.9
22.9
19.0
19.0
19.0
19.0
19.0
19.0
19.0
27.3
27.1
13.9
13.9
27.8
27.8
13.9
13.7
13.9
14.0
14.0
13.9
16.4
15.7
16.4
16.6
16.8
16.8
25.0
23.1
24.7
25.2
25.5
25.6

kips
14.9
13.7
15.2
13.9
15.2
16.5
22.0
20.4
37.9
37.9
37.9
37.9
14.8
14.8
14.8
14.8
26.5
20.1
20.6
10.3
10.3
10.3
6.2
6.2
6.2
20.9
20.9
18.8
18.8
16.8
16.8
18.2
16.8
16.8
14.6
14.6
12.4
12.4
16.1
20.9
20.8
11.5
11.5
21.8
21.8
11.5
11.3
11.5
11.6
11.6
11.5
13.0
12.4
13.1
13.1
13.3
13.4
21.1
19.5
20.9
21.3
21.5
21.6

kips
14.2
12.7
13.8
13.6
14.1
14.5
24.6
24.0
26.5*
26.8*
26.6*
25.8*
12.7*
12.7*
12.6*
12.8*
23.6
18.3
18.3
9.4
9.4
9.4
4.6
4.5
4.6
19.3
19.0
18.4
18.3
17.6
17.4
16.3
12.1
12.0
12.0
11.5
10.6
10.6
11.6
25.8
23.6
5.5
5.9
22.2
22.1
6.0
5.8
5.8
6.1
6.1
6.2
10.0
10.4
9.9
9.0
9.2
9.4
12.1
11.6
11.8
11.2
11.7
11.2

82

Table4.4Fixedfixedcolumnexperimentelasticbucklingproperties
ABAQUS elastic buckling with hole,
experiment boundary conditions

Study and Specimen Name


Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Abdel-Rahman 1997
Abdel-Rahman 1997
Abdel-Rahman 1997
Abdel-Rahman 1997
Abdel-Rahman 1997
Abdel-Rahman 1997
Abdel-Rahman 1997
Abdel-Rahman 1997
Pu et al. 1999
Pu et al. 1999
Pu et al. 1999
Pu et al. 1999
Pu et al. 1999
Pu et al. 1999
Pu et al. 1999
Pu et al. 1999
Pu et al. 1999
Sivakumaran 1987
Sivakumaran 1987
Sivakumaran 1987
Sivakumaran 1987
Sivakumaran 1987
Sivakumaran 1987
Sivakumaran 1987
Sivakumaran 1987
Sivakumaran 1987
Sivakumaran 1987
Sivakumaran 1987
Sivakumaran 1987
Sivakumaran 1987
Sivakumaran 1987
Miller & Pekoz 1994
Miller & Pekoz 1994
Miller & Pekoz 1994
Miller & Pekoz 1994
Miller & Pekoz 1994
Miller & Pekoz 1994
Miller & Pekoz 1994
Miller & Pekoz 1994
Miller & Pekoz 1994
Miller & Pekoz 1994
Miller & Pekoz 1994
Miller & Pekoz 1994
Moen and Schafer 2008
Moen and Schafer 2008
Moen and Schafer 2008
Moen and Schafer 2008
Moen and Schafer 2008
Moen and Schafer 2008
Moen and Schafer 2008
Moen and Schafer 2008
Moen and Schafer 2008
Moen and Schafer 2008
Moen and Schafer 2008
Moen and Schafer 2008

S4
S7
S6
S8
S5
S3
S14
S15
A-C
A-S
A-O
A-R
B-C
B-S
B-O
B-R
C-2.0-1-30-1
C-2.0-1-30-2
C-2.0-1-30-3
C-1.2-1-30-1
C-1.2-1-30-2
C-1.2-1-30-3
C-0.8-1-30-1
C-0.8-1-30-2
C-0.8-1-30-3
A2
A3
A4
A5
A6
A7
A8
B2
B3
B4
B5
B6
B7
B8
1-12
1-13
1-17
1-19
2-11
2-12
2-14
2-15
2-16
2-24
2-25
2-26
362-1-24-H
362-2-24-H
362-3-24-H
362-1-48-H
362-2-48-H
362-3-48-H
600-1-24-H
600-2-24-H
600-3-24-H
600-1-48-H
600-2-48-H
600-3-48-H

Boundary
Conditions
Fixed-fixed

Pcrl, LH

Pcrl, LH2

Pcrd

Pcre

Pcrl

Pcrd

Pcre

Pcrl

Pcrd

Lcrd

kips
10.7
11.9
11.6
12.5
11.4
11.2
40.5
43.5
16.2
13.7
12.3
12.9
11.2
12.1
11.8
12.3
42.6
42.5
42.6
9.5
9.5
9.5
2.8
2.8
2.8
21.3
21.5
23.5
24.8
30.5
33.4
30.2
6.0
6.1
7.0
7.6
10.3
12.4
6.6
43.2
42.5
1.7
1.7
41.3
41.3
1.8
1.7
1.8
1.9
1.9
1.8
5.9
5.4
5.7
5.3
5.2
5.7
3.3
3.2
3.5
3.4
3.4
3.4

kips
31.5
25.6
25.5
34.6
33.0
--45.5
50.1
12.8
22.4
16.4
15.8
39.5
42.0
23.0
22.7
53.0
52.9
52.9
------------------------30.8
----11.2
--20.2
19.9
11.7
51.1
50.3
2.4
2.4
41.7
41.7
2.5
2.4
2.5
2.6
2.6
2.5
6.4
5.7
6.6
5.7
5.8
6.2
3.1
2.9
3.3
3.2
3.2
3.2

kips
--41.9
33.0
--41.1
--66.9
--21.6
22.5
19.5
19.0
42.4
----29.9
85.5
85.5
85.5
48.5
48.5
48.5
17.4
17.4
17.4
------------32.2
10.4
10.6
16.5
18.6
----16.4
51.6
50.9
2.9
2.9
47.2
47.2
3.0
2.9
3.0
3.1
3.1
3.0
-------------------------

kips
40.0
43.3
41.8
42.8
44.5
41.1
86.5
70.3
24.0
24.7
23.7
24.1
45.3
45.9
30.6
30.5
109.0
109.0
109.0
50.6
50.6
50.6
17.7
17.7
17.7
57.0
58.0
50.7
50.9
81.2
81.2
81.2
12.0
12.3
19.2
19.3
21.1
21.2
18.4
76.4
75.3
3.3
3.3
74.4
74.4
3.4
3.3
3.4
3.5
3.5
3.4
9.2
10.3
9.5
9.1
9.0
9.0
7.0
6.7
7.3
5.1
5.0
5.0

kips
640.0
640.0
640.0
640.0
640.0
640.0
964.0
964.0
1014.2
1014.2
811.8
811.8
1271.6
1271.6
883.3
883.3
1182.8
1182.8
1182.8
777.4
777.4
777.4
528.3
528.3
528.3
2045.6
2045.6
2045.6
2045.6
2045.6
2045.6
1631.1
1742.4
1742.4
1742.4
1742.4
1742.4
1742.4
1742.4
1089.0
1084.1
212.7
212.7
1084.1
1084.1
231.1
231.1
231.1
231.1
231.1
231.1
119.3
112.8
130.6
30.0
29.7
36.2
239.3
238.4
242.6
56.3
53.0
55.8

kips
10.5
11.1
11.0
11.4
10.9
11.0
40.1
40.6
11.4
11.7
11.1
11.5
9.6
9.8
9.9
10.0
41.1
41.1
41.1
9.2
9.2
9.2
2.7
2.7
2.7
20.8
20.8
21.8
22.6
25.2
26.1
25.9
5.8
5.9
6.3
6.7
8.1
9.1
5.9
37.4
36.8
1.7
1.7
36.5
36.5
1.7
1.7
1.7
1.8
1.8
1.7
-------------------------

kips
17.3
17.6
17.7
17.9
17.8
18.0
38.8
40.0
14.0
14.5
13.1
13.1
22.3
23.3
20.8
20.4
52.3
52.3
52.3
19.0
19.0
19.0
11.3
11.3
11.3
35.2
35.4
38.3
39.5
35.1
36.2
33.3
10.5
10.6
12.5
13.4
16.3
15.8
16.5
35.5
35.0
2.3
2.3
35.1
35.1
2.4
2.3
2.4
2.4
2.4
2.4
-------------------------

kips
-------------------------------------------------------------------------------------------------------------------------------

kips
10.8
10.9
11.1
11.1
11.2
11.3
39.8
39.8
11.7
11.7
11.7
11.7
9.6
9.6
9.6
9.6
42.5
42.5
42.5
9.4
9.4
9.4
2.8
2.8
2.8
22.0
22.0
22.0
22.0
22.0
22.0
22.0
5.7
5.7
5.7
5.7
5.7
5.7
5.7
36.0
35.5
1.7
1.7
35.5
35.5
1.7
1.7
1.7
1.8
1.8
1.7
5.7
5.3
5.6
5.2
5.1
5.5
3.4
3.0
3.2
3.3
3.3
3.4

kips
17.7
18.0
18.0
18.0
18.1
18.1
45.5
45.4
13.8
13.8
13.8
13.8
16.9
16.9
16.9
16.9
49.9
49.9
49.9
17.2
17.2
17.2
7.5
7.5
7.5
29.4
29.4
29.4
29.4
29.4
29.4
29.4
9.8
9.8
9.8
9.8
9.8
9.8
9.8
42.1
41.6
2.1
2.1
41.6
41.6
2.1
2.1
2.1
2.1
2.1
2.1
9.5
9.0
9.3
9.0
8.9
8.9
4.9
4.9
5.0
5.1
5.1
5.1

in.
13.8
13.8
13.8
13.8
13.8
13.7
11.3
11.3
15.3
15.3
15.3
15.3
15.9
15.9
15.9
15.9
14.9
14.9
14.9
22.2
22.2
22.2
27.3
27.3
27.3
13.9
13.9
13.9
13.9
13.9
13.9
13.9
16.8
16.8
16.8
16.8
16.8
16.8
16.8
9.7
9.7
8.3
8.3
9.7
9.7
8.3
8.3
8.3
8.3
8.3
8.3
15.7
15.7
15.7
15.7
15.7
15.7
14.8
14.8
14.8
14.8
14.8
14.8

83

ABAQUS elastic buckling with hole,


CUFSM elastic buckling, no hole
CUFSM boundary conditions

Pcrl

Table4.5Weakaxispinnedcolumnexperimentdimensionsandmaterialproperties
Member
Study and Specimen Name
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981

L2
L3
L6
L7
L9
L10
L14
L16
L17
L19
L22
L26
L27
L28
L32

Boundary
Conditions
Weak-axis
pinned

Material

Hole Dimensions

nu

in.
63.0
27.0
63.0
63.0
39.0
38.9
39.1
51.0
51.1
27.0
45.0
45.0
27.0
27.0
63.0

in.
0.0490
0.0490
0.0490
0.0490
0.0490
0.0490
0.0490
0.0760
0.0760
0.0760
0.0760
0.0760
0.0760
0.0760
0.0760

ksi
29420
29420
29420
29420
29420
29420
29420
29420
29420
29420
29420
29420
29420
29420
29420

0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3

Hole Type
Circular
Circular
Circular
Circular
Circular
Circular
Circular
Circular
Circular
Circular
Circular
Circular
Circular
Circular
Circular

Cross Section Dimensions

Yield Stress and Force

h hole

r hole

B1

B2

D1

D2

Fy

P y,g

P y,net

P test

in.
0.50
1.00
1.00
1.50
1.00
1.50
0.50
1.00
1.50
1.50
1.50
1.00
1.00
1.00
1.00

in.
0.50
1.00
1.00
1.50
1.00
1.50
0.50
1.00
1.50
1.50
1.50
1.00
1.00
1.00
1.00

in.
0.25
0.50
0.50
0.75
0.50
0.75
0.25
0.50
0.75
0.75
0.75
0.50
0.50
0.50
0.50

in.
3.51
3.51
3.51
3.51
3.51
3.51
3.51
3.51
3.51
3.51
3.51
3.51
3.51
3.51
3.51

in.
1.62
1.62
1.62
1.62
1.62
1.62
1.62
1.62
1.62
1.62
1.62
1.62
1.62
1.62
1.62

in.
1.48
1.48
1.48
1.48
1.48
1.48
1.48
1.48
1.48
1.48
1.48
1.48
1.48
1.48
1.48

in.
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50

in.
0.51
0.51
0.51
0.51
0.51
0.51
0.51
0.51
0.51
0.51
0.51
0.51
0.51
0.51
0.51

in.
0.10
0.10
0.10
0.10
0.10
0.10
0.10
0.10
0.10
0.10
0.10
0.10
0.10
0.10
0.10

ksi
45.7
42.9
46.1
45.5
43.8
42.3
42.9
48.1
48.1
51.5
46.7
45.8
48.3
42.3
47.9

kips
16.2
15.2
16.3
16.1
15.5
15.0
15.2
25.9
25.9
27.7
25.1
24.7
26.0
22.8
25.8

kips
15.0
13.1
14.0
12.7
13.3
11.8
14.1
22.2
20.4
21.9
19.8
21.2
22.3
19.6
22.1

kips
8.5
11.4
8.5
8.5
9.4
10.1
9.6
17.2
15.0
21.2
20.0
19.1
21.9
22.4
13.3

Table4.6Weakaxispinnedcolumnexperimentelasticbucklingproperties
Study and Specimen Name

Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981
Ortiz-Colberg 1981

L2
L3
L6
L7
L9
L10
L14
L16
L17
L19
L22
L26
L27
L28
L32

ABAQUS elastic buckling with hole,


experiment boundary conditions
Boundary
Conditions
Weak-axis
pinned

ABAQUS elastic buckling with hole,


CUFSM elastic buckling, no hole
CUFSM boundary conditions

Pcrl

Pcrl, LH

Pcrl, LH2

Pcrd

Pcre

Pcrl

Pcrd

Pcre

Pcrl

Pcrd

Lcrd

kips
10.5
10.5
10.5
10.6
10.5
10.7
10.5
39.6
40.0
41.0
40.1
39.7
39.9
39.9
39.6

kips
16.2
13.3
12.3
12.5
14.2
12.9
16.5
39.6
41.5
41.4
42.1
39.7
39.9
43.1
39.6

kips
----------16.8
16.5
--43.9
--------45.8
---

kips
18.7
18.7
18.3
19.0
18.6
18.6
18.5
46.0
46.1
49.8
45.0
45.1
49.9
48.8
44.5

kips
8.6
30.0
8.6
8.5
19.9
18.6
19.9
18.5
18.2
50.1
23.1
23.5
52.3
56.5
12.3

kips
10.5
10.5
10.5
10.5
10.5
10.5
10.5
38.9
38.9
39.0
38.9
38.9
39.0
39.0
38.9

kips
17.9
18.9
17.9
17.9
17.7
17.2
17.7
44.6
44.6
46.6
46.1
46.3
47.5
47.5
44.5

kips
6.5
31.6
6.5
6.5
15.5
14.9
15.5
15.4
15.3
45.5
18.9
18.9
45.8
45.8
10.9

kips
10.7
10.7
10.7
10.7
10.7
10.7
10.7
39.6
39.6
39.6
39.6
39.6
39.6
39.6
39.6

kips
17.7
17.7
17.7
17.7
17.7
17.7
17.7
45.5
45.5
45.5
45.5
45.5
45.5
45.5
45.5

in.
13.8
13.8
13.8
13.8
13.8
13.8
13.8
11.3
11.3
11.3
11.3
11.3
11.3
11.3
11.3

84

Test

L hole

Table4.7Parameterrangesforfixedfixedandweakaxispinnedcolumnspecimenswithholes
hhole/h
Lhole/L
F y (ksi)
Specimen type
H/B
D/B
D/t
H/t
B/t
L/H
min
6.3 46.3 19.3
1.9
0.23
1.7
0.16
0.04
24.8
fixed-fixed columns
max 20.0 172.7 65.0
4.9
0.32 13.3
0.60
0.46
62.0
weak-axis pinned
min
6.6 46.2 20.4
2.3
0.33
7.7
0.16
0.01
42.3
columns
max 10.3 71.6 31.7
2.3
0.30 17.9
0.47
0.06
51.5

Table4.8DSMprequalificationlimitsforCsections

Column parameter

DSM
prequalification limit

Web slenderness

H/t<472

Flange slenderness

B/t<159

Lip slenderness

4<D/t<33

Web / flange

0.7<H/B<5.0

Lip / flange
Yield stress

0.05<D/B<0.41
Fy<86 ksi.

2.6.2 Boundaryconditioninfluenceonelasticbuckling
126B

TheABAQUSresultsinthecolumnelasticbucklingdatabase,inadditiontoserving

as a resource for extending DSM to columns with holes, can also be used to study the
influenceofcolumnboundaryconditionsonelasticbuckling.Considerthefixedfixed
columns in the database with L/H<4 (most are considered stub columns). Figure 4.13
146H

and Figure 4.14 and compare the influence of the experiment fixedfixed boundary
147H

conditionsforthesecolumnsrelativetowarpingfreeboundaryconditions(i.e.CUFSM
style boundary conditions in Figure 4.2) on Pcrd (distortional buckling) and Pcrl (local
148H

buckling).TheexperimentboundaryconditionsareshowntoincreasePcrdforallofthe
column specimens considered while Pcrl remains relatively unchanged, primarily
becausewarpingdeformationsareintimatelytiedtodistortionalbucklingandnotplate
buckling (Schafer and dny 2006). For stub columns, the length of the fundamental
distortionalhalfwaveisoftenshorterthanthelengthofthecolumn,whichresultsinan

85

increase in Pcrd. The restrained warping at the column ends also contributes to the
shorteningofthehalfwaveandanincreaseinPcrd.ThemagnitudeofthisboostinPcrd
decreasesasL/LcrdincreasesasshowninFigure4.13abecausethewavelengthshortening
149H

required to accommodate distortional buckling in the column can be distributed over


multiplehalfwavesascolumnlengthincreases. Figure4.13bconfirmsthisobservation
150H

bydemonstratingthatPcrdincreaseswithincreasingL/H.Hisinverselyproportionalto
Lcrd for a constant flange width B (i.e., a wider column will have a shorter distortional
halfwavelength) and therefore as L/H increases, the distortional halfwavelength
increasesrelativetothecolumnlengthcausinganincreaseinPcrd.

Pcrlincreasesslightlywithincreasingholesizerelativetocolumnsize(bothforhhole/h

andLhole/L)asshowninFigure4.14duetothefixedfixedboundaryconditions.Forlarge
15H

holes relative to member size the local buckling halfwaves form away from the hole
nearthecolumnends(seeSection3.3).Thesehalfwavelengthsareshortenedrelativeto
152H

their fundamental halfwavelengths by the loaded column edges which are also
restrained from rotating (from welding), resulting in a higher Pcrl when compared to
warpingfreeendconditionswithloadededgesfreetorotate.

86

3
Pcrd, ABAQUS experiment/Pcrd,ABAQUS warping free

Pcrd, ABAQUS experiment/Pcrd,ABAQUS warping free

2.5

1.5

0.5

0.5

1.5
L/Lcrd

2.5

2.5

1.5

0.5

0.5

1.5

2
L/H

2.5

3.5

Figure4.13InfluenceoffixedfixedboundaryconditionsversuswarpingfreeboundaryconditionsonPcrd
forcolumnexperiments(L/H<4)asafunctionof(a)columnlengthtofundamentaldistortionalhalf
wavelengthcalculatedwithCUFSMand(b)columnlengthtomemberlength.

Pcrl,ABAQUS experiment/Pcrl,ABAQUS warping free

Pcrl,ABAQUS experiment/Pcrl,ABAQUS warping free

2.5

1.5

0.5

0.1

0.2

0.3

0.4

0.5
hhole/h

0.6

0.7

0.8

0.9

2.5

1.5

0.5

0.05

0.1

0.15

0.2

0.25
Lhole/L

0.3

0.35

0.4

0.45

0.5

Figure4.14InfluenceoffixedfixedboundaryconditionsversuswarpingfreeboundaryconditionsonPcrl
forcolumnexperiments(L/H<4)asafunctionof(a)holewidthrelativetocolumnwidthand(b)holelength
relativetocolumnlength

TheweakaxispinnedboundaryconditionshaveaminimalinfluenceonPcrlandPcrd

inFigure4.15whencomparedtothewarpingfreeboundaryconditions.Thesecolumns
153H

arestillwarpingfixedeventhoughtheyarepinned(aplateisweldedtotheendofthe
member preventing warping deformation), but because the columns are all relatively
long compared to the stub columns and hole size is small relative to column size, the
wavelengthshorteningboostinPcrlandPcrdisnotpronounced.

87

3
Pcrd, ABAQUS experiment/Pcrd,ABAQUS warping free

Pcrl,ABAQUS experiment/Pcrl,ABAQUS warping free

2.5

1.5

0.5

0.05

0.1

0.15

0.2

0.25
Lhole/L

0.3

0.35

0.4

0.45

2.5

1.5

0.5

0.5

3
L/Lcrd

Figure4.15Influenceofweakaxispinnedboundaryconditionsversuswarpingfreeboundaryconditionson
(a)Pcrlasafunctionofholelengthtocolumnlengthand(b)Pcrdasafunctionofcolumnlengthtomember
length.

2.7 Approximate prediction methods for use in design


73B

The ability to approximate local, distortional, and global critical elastic buckling

loads is central to the extension of the Direct Strength Method (DSM) for coldformed
steel structural members with holes. To facilitate the use of DSM for members with
holes, approximate (and conservative) methods for calculating elastic buckling are
developedherewhichcanbeusedinlieuofafullfiniteelementeigenbucklinganalysis.
Elasticbucklingapproximationsusingthefinitestripmethod(e.g.CUFSM)arederived
for local and distortional buckling, and modifications to the classical column stability
equationsareproposedforglobalbuckling.Thesimplifiedmethodsareintendedtobe
generalenoughtoaccommodatetherangeofholeshapes,sizes,andspacingscommon
inindustry.

88

2.7.1 Localbuckling
127B

An approximate method for predicting the local elastic buckling behavior of cold

formedsteelmemberswithholesispresentedinthissection.Thismethodextendsthe
assumptionintheelementbasedmethodsin Chapter3thatlocalbucklingoccursas
154H

either plate buckling of the entire crosssection or unstiffened strip buckling at the
locationofthehole.Inthisfinitestripapproximatemethod,localbucklingisassumed
tooccursastheminimumofPcrl occurringfromlocalbucklingonthegrosscrosssection
(ascalculatedintheDirectStrengthMethod)andlocalbucklingoftheunstiffenedstrip
adjacent to the hole. The use of the finite strip method allows for a more realistic
predictionofPcrl forunstiffenedstripbucklingbyincludingtheinteractionofthecross
sectionontheunstiffenedstrip(i.e.,theLHmodefortheCsectionin Figure4.5).The
15H

method is presented through three examples considering industry standard cross


sections with holes which are then verified with ABAQUS thin shell finite element
eigenbucklingresults.Thepredictionmethodisalsovalidatedusingthecolumnelastic
bucklingdatabasedevelopedinSection4.2.6.1.
156H

4.2.7.1.1 Predictionmethod
186B

The local critical elastic buckling load Pcrl is calculated for a coldformed steel

memberwithholesas

Pcrl = min(Pcr , Pcrh ) .

(4.1)

Thecalculationofthelocalcriticalelasticbucklingloadonthegrosscrosssection,Pcr,is
performed using standard procedures defined in Appendix 1 of the AISIS100 (AISI

89

S1002007).PcrhiscalculatedinCUFSMusingthenetcrosssection,whichisrestrainedto
isolate local buckling from distortional buckling by fixing the column crosssection
cornersasshowninFigure4.16.Itisimportanttoavoidfullyrestrainingacrosssection
157H

element (i.e., flange or web), since this prevents Poissontype deformations and
artificially stiffens the crosssection. For example, Figure 4.16a restrains the corners in
158H

the zdirection only to prevent distortional buckling while still accommodating


transverse deformation of the flanges. The only time a corner should be fixed in both
the x and z directions is when two isolated elements intersect (i.e., Csection with a
flange hole, see Figure 4.16a). Finally, when a hole isolates a strip of the net cross
159H

section as shown in Figure 4.16b (e.g., a hat section with flange holes), the isolated
160H

portion of the crosssection should be deleted since it is assumed to no longer


contributestothestiffnessofthecrosssectionoverthelengthofthehole.Iftheisolated
elements are not removed then the critical elastic buckling load calculated in CUFSM
willcorrespondtoEulerbucklingofthisisolatedportionofthecrosssection.

90

Pinned (typ.)

Roller (typ.)

Web hole

Flange hole

a
b

Remove isolated
elements
Flange hole

Web hole

Figure4.16RulesformodelingacolumnnetcrosssectioninCUFSM

Oncethenetcrosssectionisrestrained,aneigenbucklinganalysisisperformed,andan
elasticbucklingcurvesimilartoFigure4.17isgenerated.Lcrhisidentifiedonthecurveas
16H

the halfwavelength corresponding to the minimum buckling load. When Lhole<Lcrh as


shown in Figure 4.17a, Pcrh is equal to the buckling load at the length of the hole. If
162H

LholeLcrh asshownin Figure4.17b,Pcrhistheminimumonthebucklingcurve.Whenno


163H

localminimumexists,thenPcrhisequaltotheelasticbucklingloadcorrespondingtoLhole.
Determining elastic buckling loads at specific halfwavelengths is a new and
fundamentally different use of the finite strip method when compared to its primary
application within DSM, which is calculating the lowest fundamental elastic buckling
modesofcoldformedsteelmembers.

91

20

20

18

Lcrh

18

16
14

14

Lhole<Lcrh

Lcrh

12
Pcr

PcrPcr
, kips

12
10

10

4
2

Lhole>Lcrh

16

Pcrh

0
0
10

Pcrh

0
1
0
10 10
half-wavelength inches
half-wavelength,

10
half-wavelength inches
half-wavelength,

Figure4.17Localelasticbucklingcurveofnetcrosssectionwhen(a)holelengthislessthanLcrhand(b)
whenholelengthisgreaterthanLcrh

4.2.7.1.2 Methodexamples
187B

Threeexamplesarepresentedherethatapproximatethelocalcriticalelasticbuckling

loadPcrlforcoldformedsteelcolumnswithholesusingCUFSM.Forallexamples,the
lengthofthecolumnL=100inchesandfiveslottedholesarespacedatS=20inches.The
typical length of the hole Lhole=4 inches. All ABAQUS eigenbuckling analyses are
modeledwithCUFSMstyleboundaryandloadingconditionsidenticaltothoseshown
in Figure4.2.Themodulusofelasticity,E,isassumedas29500ksiandPoissonsratio,
164H

,as0.3inallfinitestripandfiniteelementmodels.Pcrlisnormalizedwhenplottedby
Py,g, the squash load of the column calculated with the gross crosssectional area and a
yieldstress,Fy,of50ksi.

The first example is an SSMA 362S16233 cross section with a slotted web hole.

Figure4.18comparesthefinitestripandABAQUSmodeshapesforhhole/hC=0.14wherehC
165H

istheCsectionwebdepthmeasuredfromthecenterlineflangetocenterlineflange.The
CUFSMapproximatemethodpredictionsareplottedforarangeofhhole/hCin Figure4.19,
16H

92

andcomparedwithABAQUSeigenbucklingpredictionstodemonstratetheviabilityof
thepredictionmethod.Forthisexample,smallerholewidthsleadtoreductionsinPcrl
when compared to a member without a hole or members with larger holes. This
counterintuitive result occurs because for small holes the unstiffened strip controls the
local buckling behavior (i.e., the LH mode) and for large holes, local plate buckling
occursbetweentheholes(i.e.,theLmode),whichisconsistentwiththeelasticbuckling
observationsforplates(seeChapter3).(Onemustkeepinmindthatforstrengththenet
167H

section in yielding, as well as the elastic buckling load, ultimately determine the
capacity,notjustPcrl.)
Lcrh

hhole

hC

CUFSM Approximation
(SSMA 362S162-33)

ABAQUS

Figure4.18ComparisonofCUFSMandABAQUSpredictionsofunstiffenedstripbuckling.

93

0.5
ABAQUS
CUFSM Approx. Method

0.45
0.4
0.35

Pcrl /P y,g

0.3
0.25
0.2
0.15
0.1
0.05
0

0.1

0.2

0.3

0.4

0.5
0.6
hhole/hC

0.7

0.8

0.9

Figure4.19ABAQUSresultsverifyCUFSMlocalbucklingpredictionsforanSSMA362S16233column
withevenlyspacedwebholes.

ThenextexampleevaluatestheinfluenceofaslottedflangeholeonPcrlforanSSMA

362S16233 cross section. The unstiffened strip buckled mode shape for this cross
sectionfrombothfinitestripandfiniteelementpredictionsarecomparedinFigure4.20.
168H

It is observed that for both CUFSM and ABAQUS mode shapes, buckling occurs
primarilyinthewebandflangestrip,andthattheflangestriplipportionofthecross
section remains stable at Pcrh. The CUFSM prediction method results are plotted for
varying flange hole width bhole relative to centerline flange width bC and compared to
ABAQUSeigenbucklingpredictionsinFigure4.21.Pcrhdecreaseswithincreasingflange
169H

holewidthforbothCUFSMandABAQUSresults.Thedecreasingtrendinthecritical
elastic buckling load demonstrates the importance of the flange in web local buckling
dominatedcrosssections.

94

Assume Lcrl=Lholein CUFSM


bhole

Lcrl

bC

CUFSM Approximation
(362S162-33)

ABAQUS

Figure4.20CUFSMandABAQUSlocalbucklingmodeshapesareconsistentwhenconsideringaslotted
flangehole.
0.5
ABAQUS
0.45

CUFSM Approx. Method, Lcrl=Lhole

0.4
0.35

Pcrl /Py,g

0.3
0.25
0.2
0.15
0.1
0.05
0

0.1

0.2

0.3

0.4

0.5
0.6
bhole/bC

0.7

0.8

0.9

Figure4.21ABAQUSresultsverifyCUFSMpredictionsforanSSMA362S16233crosssectionwithevenly
spacedflangeholes.

ThethirdexampleisanSSMA1200S16268crosssectionwithaslottedholecentered

in the web. Figure 4.22 provides the CUFSM and ABAQUS buckled shapes when
170H

hhole/hC=0.16.TheassumptionintheCUFSMpredictionmethodthatLcrlisequaltoLhole=4
in.producesaPcrhhigherthanPcrwithoutthehole(becauseLcrlisshorterthanthelocal

95

buckling halfwavelength of the column) and therefore Pcr controls in the prediction
method as shown in Figure 4.23. The approximate method correctly predicts that
17H

unstiffened strip buckling does not occur as observed in the ABAQUS buckled shape,
and that the actual local buckling halfwavelength Lcrl is similar to that of a column
without holes. The prediction for Pcrl is unconservative here though (ABAQUS results
are10%lowerthanPcr),becausetheholecausesamixedlocaldistortionalmodethatis
not captured by the CUFSM netsection model (with pinned corners) or the CUFSM
grosscrosssectionmodel(withouttheinfluenceofthewebhole). Forsectionssuchas
this where local and distortional buckling have similar halfwavelengths and critical
elasticbucklingloads,afullfiniteelementeigenbucklinganalysismaybewarrantedto
evaluatethepresenceofholes.

Distortional buckling mixes


with local buckling

hhole

ABAQUS

CUFSM Approximation
(SSMA 1200S162-68)

Figure4.22ABAQUSpredictslocalplatebucklingwithdistortionalbucklinginteractionwhichisnot
detectedinCUFSM.

96

0.5
ABAQUS

0.45

CUFSM Approx. Method Lcrl=Lhole

0.4

Pcrl /Py,g

0.35
0.3

Pcrh from CUFSM is always


higher than Pcr for this
cross-section

0.25
0.2
0.15
0.1
0.05
0

Pcr

0.1

0.2

0.3

0.4

0.5

hhole/hC
Figure4.23ABAQUSresultsareslightlylowerthanCUFSMpredictions,CUFSMpredictscorrectlythat
platelocalbucklingcontrolsoverunstiffenedstripbuckling.

4.2.7.1.3 Methodvalidationusingelasticbucklingdatabase
18B

TheelasticbucklingdatabasedevelopedinSection4.2.6.1isutilizedheretoevaluate
172H

theCUFSMapproximatemethodforpredictingPcrl.CustomMatlabcodewaswrittento
calculatePcrhforall78specimensinthedatabase(Mathworks2007).Thecodeperformed
aCUFSManalysisofthenetcrosssection(crosssectioncontainingahole)withpinned
corners (x and zdirections). The predicted Pcrl of each column specimen is the
minimum of Pcrh (unstiffened strip buckling at the net section) and Pcr (Table 4.4 and
173H

Table4.6,CUFSMelasticbucklingresults,nohole).
174H

Figure 4.24 compares Pcrl reflecting the experimental boundary conditions in


175H

ABAQUS(from Table4.4andTable4.6)relativetoPcrandPcrh.Forallspecimens,Pcr(no
176H

17H

hole,grosscrosssection)islowerthanPcrh(hole,netcrosssection)becausethestripsof
webadjacenttotheholearestifferthanthecrosssectionawayfromtheholes(similarto

97

the SSMA362S16233 crosssection with hhole/h>0.20, see Figure 4.19). Even for those
178H

column specimens with small holes relative web width, the holes are often circular or
squareandthereforePcrhispredictedhigherthanPcrsincethebucklinghalfwavelength
of the unstiffened strip is assumed equal to the diameter of the hole. This prediction
resultisconsistentwiththeactualbuckledbehaviorofstiffenedelementswithcircular
andsquarecircularholesshowninFigure3.19.
179H

Figure 4.25 compares the ABAQUS experiment Pcrl to the predicted Pcrl and

180H

demonstrates the approximate method is accurate for smaller holes relative to column
sizeandbecomesincreasingconservativeasholesizeincreasesrelativetocolumnsize
(hhole/h and Lhole/L). The prediction becomes conservative because it does not take into
accountthewavelengthstiffeningeffects(discussedinSection 3.3.2)whichboostPcrlas
18H

theholebecomeslargerelativetothecolumn.Themeanandstandarddeviationofthe
ABAQUStopredictedratioare1.11and0.18respectively,demonstratingtheviabilityof
themethodforthespecimensconsidered.
3

3
Pcrh (CUFSM net section)

1.5

0.5

Pcr (CUFSM gross section)

2.5

Pcrl,ABAQUS/Pcrl,predicted

2.5

Pcrl,ABAQUS/Pcrl,predicted

Pcrh (CUFSM net section)

Pcr (CUFSM gross section)

1.5

0.5

0.1

0.2

0.3

0.4

0.5
hhole/h

0.6

0.7

0.8

0.9

0.1

0.2

0.3

0.4

0.5
Lhole/L

0.6

0.7

0.8

0.9

Figure4.24PredictedPcrh(CUFSM,bucklingofthenetcrosssection)andPcr(CUFSM,bucklingofthegross
crosssection,nohole)arecomparedrelativetotheABAQUSPcrlwithexperimentboundaryconditionsasa
functionof(a)holewidthtoflatwebwidthand(b)holelengthtocolumnlength

98

2.5

2.5

Pcrl,ABAQUS/Pcrl,predicted

Pcrl,ABAQUS/Pcrl,predicted

1.5

0.5

1.5

0.5

0.1

0.2

0.3

0.4

0.5
hhole/h

0.6

0.7

0.8

0.9

0.1

0.2

0.3

0.4

0.5
Lhole/L

0.6

0.7

0.8

0.9

Figure4.25PredictedPcrl(CUFSMapproximatemethod)iscomparedrelativetotheABAQUSPcrlwith
experimentboundaryconditionsasafunctionof(a)holewidthtoflatwebwidthand(b)holelengthto
columnlength

2.7.2 Distortionalbuckling
128B

An approximate method utilizing the finite strip method is developed here for

predictingthedistortionalcriticalelasticbucklingloadPcrdofcoldformedsteelcolumns
withholes.Themethodsimulatesthelossinbendingstiffnessofacrosssectionwithin
a distortional buckling halfwave by modifying the crosssection thickness in CUFSM.
Two different approaches to simulating this loss in stiffness are evaluated. The first
approach reduces the member thickness in the regions of the crosssection with holes
basedontheratioofholelengthtodistortionalhalfwavelength.Thesecondapproach
isdevelopedforCsectionswithwebholesandismechanicsbased,wherethethickness
of the entire web is reduced based on an assumed relationship between web bending
stiffness (derived with observations from ABAQUS thin shell elastic FE analyses) and
thebendingstiffnessmatrixtermsofafinitestripelement.Thestepsforimplementing
these methods in CUFSM are described, and an example is provided where the
predictionmethodforPcrdiscomparedtoABAQUSeigenbucklingresultsforanindustry

99

standardSSMA250S16268crosssectionwithevenlyspacedwebholesalongthelength
of a column. An empirical equation is derived to account for the increase in Pcrd from
warping fixed end conditions and then the viability of the approximate method is
evaluatedagainstPcrdfromthecolumnexperimentdatabaseinSection4.2.6.1.
182H

4.2.7.2.1 Predictionmethod
189B

The prediction method presented here for Pcrd assumes that the change in cross

sectional stiffness within a distortional halfwave caused by the presence of a hole (or
holes) can be simulated by assuming a reduced thickness of the crosssection. The
distortional halfwavelength of the crosssection, Lcrd, without holes is determined first.
TheelasticbucklingcurveiscalculatedusingthegrosssectionofthecolumninCUFSM
and Lcrd is read off of the curve at the location of the distortional local minimum as
shown in Figure 4.26 (this elastic buckling curve corresponds to an SSMA 250S16268
183H

crosssection,whereLcrd=12in.).ThepredictionmethodassumesthatLcrddoesnotchange
with the presence of holes. The crosssection is then modified to approximate the
presence of holes within a distortional halfwavelength. Two approaches for this
modificationsteparepresentednextinSection4.2.7.2.1.1andSection4.2.7.2.1.2. Once
184H

185H

the cross section is modified to account for the presence of a hole in CUFSM, another
elastic buckling curve is generated and Pcrd (including the presence of the hole) is
determinedastheelasticbucklingloadoccurringatLcrdasshowninFigure4.26.
186H

100

100
without hole
with hole

90
80
70

Lcrd (determined at
local minimum of no
hole curve)

P cr , kips

60
50
40
30
20
10
0
0
10

Pcrd (includes
influence of hole)

10
half-wavelength, in.

10

Figure4.26CUFSMapproximatemethodforcalculatingPcrdforacolumnwithholes.

.7.2.1.1 Weighted average approach for predicting hole influence


209B

Theholeinfluenceondistortionalbucklingofanopenthinwalledcrosssectioncan

beapproximatedbymodifyingthecrosssectionthicknessinCUFSMatthelocationofa
holewiththefollowingequation:

L
t hole = 1 hole t .
Lcrd

(4.2)

The implementation of the reduced thickness in a Csection with a single web hole is
provided in Figure 4.27. This approach is an intuitive first cut at approximating the
187H

reduction in bending stiffness of the cross section. A more refined mechanicsbased


approachispresentednext.

101

thole

hhole

Figure4.27ModifiedcrosssectiontobeusedinCUFSMtopredictPcrdforacolumnwithholes.

.7.2.1.2 Mechanics-based approach for predicting hole influence


210B

A plate model is developed in ABAQUS to study the influence of a hole on the

bendingstiffnessofaSSMA250S16268columnexperiencingdistortionalbuckling.The
stiffness reduction observed in ABAQUS is quantified and then equated to finite strip
bending stiffness matrix terms to derive a reduced web thickness expression based on
finite strip mechanics. The plate dimensions in ABAQUS are chosen to correspond to
thewebofthe250S16268sectionoveronedistortionalhalfwave.Theplatewidthhis
2.4 in., the plate length L=12 inches (consistent with Lcrd=12 in.), and t=0.0713 in. One
slotted hole with Lhole=4 in. is centered in the plate. The width of the hole is varied,
hhole=0.5 in., 0.96 in., 1.20 in., 1.5 in., and 1.75 in. (and subsequently rhole varies). The
modulusofelasticity,E,isassumedas29500ksiandPoissonsratio,,as0.3forallfinite
element models considered here. The ABAQUS boundary conditions and applied
loading are described in Figure 4.28. The plate is simplysupported and loaded with
18H

imposedrotationsatthelongedgesoftheplatewithmagnitudesvaryingasahalfsine
wavetosimulatedistortionaldeformationoveronehalfwavelength.

102

Restrain midline node in


1 and 3

1 (x)

Restrain midline node in 3


Apply imposed rotation at
plate edges in the shape of
0.001*sin(x/L) radians
Restrain perimeter in 2

Figure4.28ABAQUSboundaryconditionsandimposedrotationsforwebplate

Thedeformedshapeoftheplatewhenhhole/h=0.50isprovidedinFigure4.29.Ateach
189H

node where an imposed rotation is applied, the associated moment is obtained from
ABAQUS and plotted in Figure 4.30 as a transverse bending stiffness per unit length.
190H

(Notethatnearx=0in.andx=12in.,thedeformedshapeinABAQUSresultsinasmall
negative bending stiffness which is not plotted in Figure 4.30 and does not affect the
19H

overallresultshere.Thenegativestiffnessisnotpredictedinthefinitestripformulation
because the longitudinal shape function is enforced as a halfsine wave). The hole
causes a sharp reduction in bending stiffness at the location of the hole, but has a
minimal influence on bending stiffness away from the hole. The stiffness reduction is
shown to be relatively insensitive to the ratio of hole width to plate width except for
peaksinstiffnessthatincreasewithhhole/hattheroundededgesoftheslottedhole.The
results in Figure 4.30 confirm the intuitive assumption employed to develop Eq. (4.2);
192H

193H

the ratio of the length of the hole to the length of the distortional halfwave is an
importantparameterwhenpredictingthelossinbendingstiffness.

103


Figure4.29Platedeformationfromimposededgerotations,hhole/h=0.50

0.07
hhole/h=0.63

rotational stiffness, k*in/rad

0.06

0.05
no hole

0.04

0.03

0.02
Increasing
hhole/h

0.01

4
6
8
distance along plate, in.

10

12

Figure4.30Transverserotationalstiffnessoftheplateissignificantlyreducedinthevicinityoftheslotted
hole

If K represents the cumulative transverse bending stiffness for the plate without a

hole(areaunderthecurvein Figure4.30),thenthereducedKincludingthepresenceof
194H

theholecanbeapproximatedas:

L
K hole = 1 hole K .
Lcrd

104

(4.3)

TheglobalbendingstiffnessKforasimplysupportedfinitestripelementisderivedby
applyingaunitrotationatthestripedges:

keb [d w ]

k11
k
= 12
k31

k 41

k12
k 22

k13
k 23

k32
k 42

k33
k 43

k14 0 V
k 24 1 K
= ,
k34 0 V

k 44 1 K

(4.4)

where the keb is the bending stiffness matrix and dw=[w1 1 w2 2] (Schafer and dny
2006).SolvingEq.(4.4)forK:
195H

K = k 22 k 24 .

(4.5)

Sincek22andk24arebothfunctionsofthewebthickness(tweb)3,KandKholecanbeequated
directlyas:

K hole t web,hole
=
.
3
K
t web

( )

(4.6)

Substituting Eq. (4.6) and rearranging in terms of tweb,hole, the reduced web thickness
196H

correspondingtothereducedtransversebendingstiffnessfromtheholeis:

t web,hole

L
= 1 hole
Lcrd

1/ 3

t web .

(4.7)

Eq. (4.7) is an improvement over Eq. (4.2) because it reflects the underlying plate
197H

198H

bending mechanics involved in distortional buckling and is actually simpler to


implement in CUFSM since the entire web thickness of a Csection is reduced to tweb,hole
insteadofchangingthesheetthicknessjustatthelocationoftheholeasshowninFigure
19H

4.27.Asimilarmodificationtothasbeenproposedforwebslottedthermalstructural
studs(Kesti2000).

105

4.2.7.2.2 Methodexample
190B

ThedistortionalcriticalelasticbucklingloadPcrdiscalculatedherewiththeCUFSM

predictionmethodforalongcolumn(L=100in.)withanSSMA250S16268crosssection
andfiveevenlyspacedslottedwebholes(S=20in.,Lhole=4in.).Thewidthoftheholeis
varied relative to the web width, and ABAQUS eigenbuckling results are used to
evaluatetheviabilityofthemethod.AllABAQUSfiniteelementmodelshaveCUFSM
style boundary and loading conditions as shown in Figure 4.2. The modulus of
120H

elasticity,E,isassumedas29500ksiandPoissonsratio,,as0.3inallfinitestripand
finiteelementmodels.PcrdisnormalizedbyPy,gwhenplotted.Py,gisthesquashloadof
thecolumncalculatedwiththegrosscrosssectionalareaandassumingFy=50ksi.

AcomparisonoftheCUFSMpredictionmethod(employingtheweightedaverage

thicknessapproximation)andABAQUSdistortionalbucklingmodeshapesareprovided
in Figure4.31whenhhole/h=0.63.Ninedistortionalhalfwavesformalongthememberin
120H

ABAQUS, with every other halfwave containing one slotted hole. The CUFSM
prediction method employing both the weighted average and mechanicsbased
thickness modifications to the crosssection are compared over a range of hhole/h to
ABAQUS eigenbuckling results in Figure 4.32. Pcrd for the pure ABAQUS distortional
120H

(D) buckling mode is plotted to demonstrate that prediction method is viable for this
crosssection and hole spacing. The CUFSM prediction for Pcrd with the weighted
averagethicknessreductionattheholedecreaseswithincreasingholewidthsincethe
webprovideslessbendingstiffnesstotheflangesasmoreholematerialisremoved.The
CUFSMpredictionemployingthemechanicsbasedreductioninwebthicknessisnot

106

afunctionofhhole/handisshowntobeamorerealisticpredictorofPcrdthantheweighted
average approach. These approximate methods are evaluated against ABAQUS
distortionalbucklingpredictionsfromthecolumndatabaseinSection4.2.7.2.4.
1203H

Lcrd
hhole

CUFSM Approximation
(250S162-68)

ABAQUS

Figure4.31ComparisonofCUFSMandABAQUSdistortionalbucklingmodeshapes.

3
CUFSM Approx. Method with "weighted average" t hole
CUFSM Approx. Method with "mechanics-based" t w eb,hole

2.5

ABAQUS D mode

Pcrd/Py,g

1.5

0.5

0.1

0.2

0.3

0.4

0.5
hhole/h

0.6

0.7

0.8

0.9

Figure4.32CUFSMdistortionalbucklingpredictionmethodisconservativewhenconsideringanSSMA
262S16268columnwithuniformlyspacedholes.

107

4.2.7.2.3 Warpingfixeddistortionalamplificationfactor
19B

Longitudinalwarpingdeformationsoccurasaresultofdistortionalbucklingincold

formed steel columns. When this warping deformation is restrained, the distortional
buckling halfwavelength is shortened relative to the fundamental distortional half
wavelengthofthecolumncrosssection,Lcrd.Thischangeinhalfwavelengthresultsin
anamplificationofthedistortionalcriticalelasticbuckling loadofthecolumn,Pcrd(see
Figure4.13aforboostinPcrdforstubcolumns).Theelasticbucklingdatabasedeveloped
1204H

in Section 4.2.6.1 (Table 4.4 and Table 4.6) provides an opportunity to derive an
1205H

1206H

1207H

empiricalamplificationfactorsinceforallcolumnsinthedatabase,Pcrdforbothwarping
fixed (ABAQUS) and warping free (CUFSM) boundary conditions are known and the
fundamentaldistortionalhalfwavelength,Lcrd,hasbeencalculatedinCUFSM.

ThewarpingfixedboundaryconditioneffectonPcrdisplottedforthe78specimensin

thecolumndatabaseinFigure4.33.TheboostinPcrdishighestwhenthecolumnisshort
1208H

relativetoLcrdbecausethewavelengthshorteningmustbeaccommodatedoverjustone
halfwave.Anempiricalequation(alsoplottedinFigure4.33)isfittothedatatrend:
1209H

D boost

1L
= 1 + crd
2 L

(4.8)

ThisamplificationfactorcanbeusedwiththeCUFSMpredictionmethoddevelopedin
Section 4.2.7.2.1 when the column being considered has warpingfixed boundary
120H

conditions.Eq.(4.8)isconsistentwiththedistortionalbucklingboostfactorprovidedin
12H

theDSMDesignGuide(AISI2006)asshowninFigure4.33.
12H

108

3
Specimen data
Eq.(4.8)
DSM Design Guide

Pcrd,fixed-fixed /Pcrd,CUFSM warping free

2.5

1.5

0.5

0.5

1.5

2.5
L/Lcrd

3.5

4.5

Figure4.33WarpingfixedboundaryconditionamplificationofPcrd

4.2.7.2.4 VerificationofCUFSMapproximatemethodwithcolumndatabase
192B

The CUFSM approximate method for distortional buckling is now evaluated using

theelasticbucklingpropertiesofthe78columnspecimensfromtheexperimentdatabase
developed in 4.2.6.1. The ABAQUS distortional critical elastic buckling load Pcrd,
123H

determined with the experiment boundary conditions, is plotted against the


approximate method predictions in Figure 4.34. The predictions including the
124H

distortionalamplificationfactorfromEq.(4.8).Theapproximatemethodemployingthe
125H

weighted average reduced web thickness at the hole from Eq. (4.2) and the
126H

mechanicsbasedreducedwebthicknessapproachfromEq. (4.7)arebothpresented.
127H

Theaccuracyofthepredictionmethodimprovesasthecolumnlengthincreasesrelative
tothefundamentaldistortionalhalfwavelengthLcrd.Thepredictionaccuracyishighly

109

variable when L/Lcrd<1, primarily because of the variation in the boundary condition
influence described in Section 4.2.7.2.3 for stocky columns. As expected, the
128H

mechanicsbased thickness approach (with ABAQUS to predicted ratio mean and


standard deviation of 1.19 and 0.29) is more accurate over the 78 columns than the
weightedaverageapproach(meanof1.24andstandarddeviationof0.29).

3
weighted average t hole
mechanics-based t w eb,hole

Pcrd,ABAQUS Exp. BC/Pcrd,predicted

2.5

1.5

0.5

3
L/Lcrd

Figure4.34AccuracyoftheCUFSMapproximatemethodforpredictingPcrdimprovesascolumnlength
increasesrelativetothefundamentaldistortionalhalfwavelengthforwarpingfixedcolumns

2.7.3 Globalbuckling
129B

Twodifferentapproximatemethodsforcalculatingthecriticalelasticbucklingload

ofacolumn,Pcre,areevaluatedinthissection.Bothmethodsemployweightedaverages
ofthemembersectionpropertiesintheclassicalcolumnstabilityequationtoaccountfor
theinfluenceofholes,oneusingaweightedcrosssectionalthicknessatthelocationsof
theholeandtheotherusingtheweightedaverageofthegrossandnetcrosssections.The

110

approximatemethodsarecomparedtoABAQUSeigenbucklingresultsforalongcold
formed steel column (SSMA 1200S16268 cross section) with uniformly spaced circular
holes.Theaveragetorsionalconstants,JandCw,arecalculateddirectlyusingABAQUS
for this column and then compared to their associated weighted average estimates.
Basedonthesestudies,recommendationsaremaderegardingtheapproximatemethod
mostsuitableforpredictingPcreforcolumnswithholes.

4.2.7.3.1 Descriptionofmethods
193B

.7.3.1.1 Weighted Properties Method


21B

Theequationforpredictingtheglobal(flexuralonly)criticalelasticbucklingloadPcre

of a column with holes along its length can be solved using energy methods, and is
derived for a column with two holes located symmetrically about the longitudinal
midlineofthecolumnin AppendixE.TheequationthatevolvesfromtheRaleighRitz
129H

derivationforthiscaseis:

Pcre =

2 E I g L NH + I net LH

L2

(4.9)

whereIgisthemomentofinertiaofthegrosscrosssection,Inetisthemomentofinertiaof
thenetcrosssection,LNHisthelengthofcolumnwithoutholesandLHisthelengthof
columnwithholes(notethatLNH+LH=L).Theaveragemomentofinertiaofthecolumn
with holes is shown in Eq. (4.9) to be equivalent to the weighted average of the gross
120H

andnetcrosssectionsalongthecolumnlength.Anapproximatemethodforcalculating
PcreisproposedherewhichextendsthisweightedpropertiesmethodologyinEq. (4.9)
12H

111

toallofthecrosssectionpropertiesofthecolumnrequiredtosolvetheclassical cubic
bucklingequationforcolumns(Chajes1974):

(P

cre , y

P )(Pcre, x P )(Pcre, P ) (Pcre, y

P 2 xo2
P 2 y o2
P ) 2 (Pcre, x P ) 2 = 0
ro
ro

(4.10)

including the crosssectional area A, moment of inertia Ix and Iy, St. Venant torsional
constant J, and shear center location. The computer program CUTWP solves Eq. (4.10)
12H

for any general crosssection and is freely available (Sarawit 2006). The net section
properties can be calculated in CUFSM (or CUTWP) by reducing the sheet strip
thickness to zero at the location of the hole. This approximation is referred to as the
weightedpropertiesmethod.AgeneralformofEq. (4.9)isalsoderivedin Appendix
123H

124H

E which can be used with the weighted properties method for the case of a column
withasingleholeormultiplearbitrarilyspacedholes.

.7.3.1.2 Weighted Thickness Method


21B

Thisapproximatemethodapproachesthecalculationoftheaveragecolumnsection

propertiesinadifferentway,byusingaweightedmemberthicknessatthelocationof
theholesinthecrosssectiontocalculatetheaveragecrosssectionalproperties:

t hole =

L LH
t .
L

(4.11)

An example of a Csection with a reduced thickness at the location of a web hole is


provided in Figure 4.35. This weighted thickness method is more convenient to
125H

implement than the weighted properties method presented in the previous section

112

becausethemodifiedcrosssection(withreducedthickness)canbeinputdirectlyintoa
computerprogramsuchasCUTWP.

hhole

thole

Figure4.35Aweightedthicknesscrosssectioncanbeinputdirectlyintoaprogramthatsolvesthe
classicalcubicstabilityequationforcolumns(e.g.CUTWP).

4.2.7.3.2 Exampleandverification
194B

ABAQUSglobaleigenbucklingresultsarecomparedinthissectiontotheweighted

properties and weighted thickness prediction methods for an industry standard


SSMA 1200S16268 long column (SSMA 2001) with evenly spaced circular holes. The
lengthofthecolumnL=100in.,theholespacingS=20in.,andthediameterofthecircular
holeisvariedfromhhole/H=0.10tohhole/H=0.90whereHistheouttooutdepthofthecross
section (see Figure 4.1 for crosssection dimension notation). All ABAQUS finite
126H

elementmodelsareloadedincompressionatthememberendsandhavewarpingfree
boundaryconditionsasshowninFigure4.2.Themodulusofelasticity,E,isassumedas
127H

29500ksiandPoissonsratio,,as0.3inallCUTWPandfiniteelementmodels.

The three global buckling modes of this SSMA 1200S16268 long column without

holesarecalculatedinCUTWP:(1)weakaxisflexuralbucklingoccursatPcr=7.91kips,
(2) flexuraltorsional buckling occurs at Pcr=13.39 kips, and (3) strongaxis flexural

113

buckling occurs at Pcr=604.17 kips. The first two buckling modes are the focus of this
study since the strongaxis buckling mode is much higher than the squash load of the
column (Py,g= 56.30 kips assuming Fy=50 ksi) and will not influence the design of the
column. Figure 4.36 provides an example of the weakaxis flexural and flexural
128H

torsional buckling modes when hhole/H=0.50. Note that shell FE predicts local buckling
mixingwiththeweakaxisflexuralmodewhenhhole/H>0.50becausePcreisreducedbythe
presence of holes to a magnitude similar to the local critical elastic buckling load
(Pcrl=6.69kips). Local buckling is not observed to mix with global buckling in the
flexuraltorsional(column)orlateraltorsional(beam)bucklinginthisstudy.
Web local buckling
mixes with global mode

Weak Axis Flexural


Pcre=6.96 kips

Flexural-Torsional
Pcre=10.64 kips

Global column buckling modes


(SSMA 1200S162-68, hhole/h=0.50)

Figure4.36WeakaxisflexuralandflexuraltorsionalglobalbucklingmodesforanSSMA1200S16268
columnwithevenlyspacedcircularholes.

.7.3.2.1 Section property calculations at the net section


213B

To draw meaningful conclusions regarding the weighted properties and

weighted thickness prediction methods it is first helpful to understand how circular


holediameterinfluencesthecolumnssectionproperties. Figure4.37comparesthenet
129H

sectiontothegrosscrosssectionalareaA,thestrongandweakaxismomentofinertiaIx
andIy,theSt.VenanttorsionalconstantJ,andthewarpingtorsionalconstantCwofthe

114

columnastheyvarywithhhole/H.Allnetsectionpropertiesinthisfigurearedetermined
withtheCUFSMsectionpropertycalculatorbyreducingthesheetthicknesstozeroat
the location of the hole. A and J decrease linearly with hole diameter while Ix and Iy
decreasenonlinearly.Iyismostsensitivetothepresenceoftheholebecausetheholeis
locatedinthechannelwebforthiscase,whichprovidesmuchofthecontributiontothe
weakaxismomentofinertia.Cw,calculatedhereassumingzerothicknessbutcontinuity
through the hole, varies nonlinearly with hhole/H. It is unclear if the net section Cw
calculatedinthiswayproducesthebestapproximationofthecolumnsactualphysical
behavior in torsional buckling. The magnitude of Cw is influenced heavily by cross
sectioncontinuitysince alineintegralaroundthecrosssectionisusedtosolveforthe
warpingfunction.FurtherinvestigationofJandCwforcolumnswithholesispresented
inSection4.2.7.3.2.3.
1230H

1.5
Cw
Ix (strong)

net section/gross section

Iy (weak)
A
J

0.5

0.1

0.2

0.3

0.4

0.5
hhole/H

0.6

0.7

0.8

0.9

Figure4.37Variationinnetsectionpropertiesascircularholediameterincreases.

115

.7.3.2.2 Average section property calculations for the column - A, Ix, and Iy
214B

The average section properties of the 1200S16268 column with circular holes

calculated using the weighted thickness and weighted properties methods are
compared in Figure 4.38 through Figure 4.40. For this example problem there are
123H

123H

minimal differences between the methods for A and Ix, although Iy calculated with the
weightedpropertiesmethoddecreasesinstiffnessrelativetotheweightedthickness
methodasholediameterincreasesrelativetocolumnwidth.

1.5
weighted thickness
weighted properties

Aavg/Ag

0.5

0.1

0.2

0.3

0.4

0.5
hhole/H

0.6

0.7

0.8

0.9

Figure4.38Comparisonofweightedthicknessandweightedpropertiescrosssectionalarea.

116

1.5

Ix,avg/Ix,g (strong axis)

weighted thickness
weighted properties

0.5

0.1

0.2

0.3

0.4

0.5
hhole/H

0.6

0.7

0.8

0.9

Figure4.39Comparisonofweightedthicknessandweightedpropertiesstrongaxismomentofinertia.

1.5

Iy,avg/Iy,g (weak axis)

weighted thickness
weighted properties

0.5

0.1

0.2

0.3

0.4

0.5
hhole/H

0.6

0.7

0.8

0.9

Figure4.40Comparisonofweightedthicknessandweightedpropertiesweakaxismomentofinertia.

.7.3.2.3 Average section property calculations for the column - J and Cw


215B

The average J and Cw of the 1200S16268 column with circular holes is determined

directly using ABAQUS and then compared to the weighted properties and

117

weightedthicknesspredictionsinthissection.Thedifferentialequationfortorsionis
definedas(Timoshenko1961):

d
d 3
T = GJ
ECw 3 ,
dx
dx

(4.12)

where is the angle of twist of the cross section and G is the shear modulus of steel
(G=11346ksiinthiscase).Eq.(4.12)isusedinconjunctionwithstaticABAQUSanalyses
123H

(noteigenbuckling!)tosolveforJavgandCw,avgofthecolumnashhole/Hvariesfrom0.10to
0.90.Javg iscalculatedbyapplyingaunittwistattheendofthecolumnaboutthegross
crosssectionshearcenterwhilekeepingtheoppositecolumnendfixedagainsttwistas
shownin Figure4.41.Ifbothendsofthecolumnarefreetowarp,thevariationintwist
1234H

alongthecolumnis constantasshownin Figure4.42andwarpingresistancedoesnot


1235H

contributetotheresultingtorsion(d3/dx3=0).Thevariationintwistwasnotsensitiveto
increasingholediameterinthiscase,andthereforethelineshownin Figure4.42isthe
1236H

same regardless of hole diameter. The twist along the column is measured in
ABAQUSastherelativerotationoftheflangewebcorners.Thetwistmagnitudealong
the column length remains unchanged with increasing hhole/H. Javg for the warping free
columniscalculatedbyrearrangingEq.(4.12):
1237H

J avg =

To L

G o

(4.13)

o,G,andLareknownandToisthetorqueresultingfromtheunitrotationo,whichis
readdirectlyfromABAQUS.

118

shear center reference


node (determined with
gross cross section)

Cross-section kinematically restrained to shear


center reference node in 2, 3 (warping free)

SECTION A-A

1 (x)

Cross-section dof
fixed in 2 and 3
(warping free)
Node at centerline
of web fixed in 1 to
prevent rigid body
motion

Figure4.41ABAQUSboundaryconditionsforwarpingfreeandappliedunittwistatx=0in.andwarping
freebutrotationrestrainedatx=100in.

1.4

/0, normalized angle of twist

1.2

0.8

0.6

0.4

0.2

10

20

30

40
50
60
70
distance along column, in.

80

90

100

Figure4.42AngleoftwistdecreaseslinearlyintheSSMA1200S16268columnwithwarpingfreeend
conditions.

The resulting Javg from ABAQUS is compared against the weighted properties and
weighted thickness calculations of Javg. (Note that the weighted properties Javg is
calculated with Jnet from Figure 4.37 using the CUFSM section property calculator and
1238H

119

assuming the thickness is zero at the hole). It is clear from Figure 4.43 that the
1239H

weighted properties calculation of Javg is most consistent with Javg derived from
ABAQUS.

1.5
weighted thickness
weighted properties
ABAQUS

Javg/Jg

0.5

0.1

0.2

0.3

0.4

0.5
hhole/H

0.6

0.7

0.8

0.9

Figure4.43TheweightedpropertiesapproximationforJavgmatchescloselywiththeABAQUSprediction
fortheSSMA12S0016268columnwithholes

The ABAQUS boundary conditions are now modified such that warping is

restrained at the fixed column end as shown in Figure 4.44. A unit twist, o, is again
1240H

appliedatthefreeend,andtheresultingangleoftwistalongthelengthofthecolumn
is measured in ABAQUS. Because of the warpingfixed end condition, is nonlinear
along the length of the column and warping resistance contributes to the torsional
stiffnessofthecolumn.Sincethedistributionofalongthecolumnisnotinfluencedby
hhole/H as observed in ABAQUS, an indirect solution of Cw,avg as a function of Cw,g can be
derived:

120

C w,avg

C w, g

d
(x = 0)
dx
,
d
(x = 0)
GJ g
dx

To GJ avg
To , g

(4.14)

where for each ABAQUS model (hhole/H=0.10 to 0.90), the torque To associated with the
unittwistoisreaddirectlyfromABAQUSandd/dx(x=0)iscalculatedfromFigure4.45.
124H

Aswasthecaseforthewarpingfreecasein Figure4.42,thevariationintwistwasnot
124H

sensitivetoincreasingholediameterandthereforethelineshownin Figure4.45isthe
1243H

same regardless of hole diameter. The influence of holes on the variation in twist is
expected to be more pronounced as column length decreases relative to hole length.
Futureresearchisplannedtoevaluatetheinfluenceofmemberlengthonthetorsional
propertiesofcolumnswithholes.

Cross-section kinematically restrained to shear


center reference node in 2, 3 (warping free)
shear center reference
node (determined with
gross cross section)

SECTION A-A

1 (x)

Cross-section dof
fixed in 1,2, and 3
(warping fixed)

Figure4.44ABAQUSboundaryconditionsforwarpingfreeandappliedunittwistatx=0in.andwarping
fixedandrotationrestrainedatx=100in.

121

1.4

/0, normalized angle of twist

1.2

0.8

0.6

0.4

0.2

10

20

30

40
50
60
70
distance along column, in.

80

90

100

Figure4.45AngleoftwistisnonlinearalongtheSSMA1200S16268columnwithwarpingfixedend
conditionsatx=100in.

Figure 4.46 demonstrates that the weighted properties and weighted thickness
124H

approximations both overestimate Cw,avg when compared to the ABAQUS derived Cw,avg
demonstrating that neither is an accurate predictor of Cw,avg. A modified version of the
weightedpropertiesapproximationisalsoplotted,whereCw,netistakenequaltozero
instead of Cw,net taken from the results in Figure 4.37. This assumption for Cw,net is
1245H

motivated by the idea that the hole separates the cross section into two pieces, where
each piece on its own contributes minimally to warping resistance. This modified
weighted properties approximation results in a conservative lower bound on Cw,avg
which is useful from a design perspective until more accurate approximations are
developed.

122

1.5
weighted thickness
weighted properties
derived from ABAQUS
weighted properties Cw ,net=0

Cw,avg /Cw,g

0.5

0.1

0.2

0.3

0.4

0.5
hhole/H

0.6

0.7

0.8

0.9

Figure4.46Comparisonofweightedthicknessandweightedpropertiesapproximationstothe
ABAQUSderivedwarpingtorsionconstantCw,avg.

.7.3.2.4 Comparison of prediction accuracy between methods


216B

Figure 4.47 compares the weakaxis flexural critical elastic buckling load of the
1246H

1200S16268 column calculated with the weighted thickness and weighted


properties prediction methods to ABAQUS eigenbuckling results. The ABAQUS
calculation of Pcre is systematically 10% lower than the prediction method (even for a
columnwithoutholes),whichresultsfromtheassumptionofarigidcrosssectioninthe
classical stability equations. The column crosssection as modeled in ABAQUS is
allowed to change shape along the length, resulting in a lower axial stiffness of the
column. (The reduction in Pcre was also confirmed in CUFSM, which like ABAQUS,
accounts for platetype deformations in elastic buckling calculations.) Beyond this
systematic difference, both approximate methods are accurate predictors of Pcre for

123

hhole/h0.60andtheweightedpropertiesmethodremainsaccurateforevenlargerholes.
ThepredictionofweakaxisflexurePcreusingthenetsectionpropertiesfrom Figure4.37
1247H

isalsoplottedin Figure4.47,demonstratingaconservativealternativetotheweighted
1248H

propertiesandweightedthicknessmethods.

1.5

Pcre,ABAQUS/Pcre,prediction

weighted thickness
weighted properties
net section
1

Difference is caused by classical rigid


cross-section assumption which
increases axial column stiffness
when compared to ABAQUS

0.5

0.1

0.2

0.3

0.4

0.5
hhole/H

0.6

0.7

0.8

0.9

Figure4.47Comparisonofweightedthicknessandweightedpropertiespredictionmethodsforthe
SSMA1200S16268weakaxisflexuralbucklingmode.Predictionsusingnetsectionpropertiesarealso
plottedasaconservativebenchmark.

Figure4.48comparestheweightedthicknessandweightedpropertiesmethods
1249H

toABAQUSresultsforthesecondglobalmode,flexuraltorsionalcolumnbuckling.The
accuracy of the prediction methods decrease with increase hhole/H for both methods,
confirmingwhatwasobservedin Figure4.46,thattheweightedapproximationsforCw
1250H

are not accurate representations of the average warping torsion stiffness, especially as
hhole/Hbecomeslarge.WarpingtorsiondominatesoverSt.Venanttorsioninthismode
since both weighted average methods predict similar variations in Pcre, even though J
variesbetweenthetwomethods(see Figure4.43).Theweightedpropertiesmethod
125H

124

withCw,avgreplacedwithCw,avgpredictedinABAQUS(see Figure4.46)accuratelypredicts
125H

Pcre until hhole/H exceeds 0.80, although this method may not be practical from a design
perspective since it requires thin shell FE analysis. The modified weight properties
approach,calculatedassumingCw,net=0,isshowntobemoreaccuratethanusingjustthe
net section properties and is a conservative method for predicting Pcre of flexural
torsional buckling modes. Future research is planned to develop a mechanicsbased
approximationfortheaverageCwofacolumnincludingtheinfluenceofholes.

Pcre,ABAQUS/Pcre,prediction

1.5

weighted thickness
weighted properties
net section
weighted properties, ABAQUS Cw ,avg

0.5

weighted properties, Cw ,net=0

0.1

0.2

0.3

0.4

0.5
hhole/H

0.6

0.7

0.8

0.9

Figure4.48Comparisonofweightedthicknessandweightedpropertiespredictionmethodsforthe
SSMA1200S16268flexuraltorsionalcolumnbucklingmode.Predictionsusingnetsectionpropertiesare
alsoplottedasaconservativebenchmark.

125

4.3 Elasticbucklingofbeamswithholes
3B

3.1 Analysis of existing tests on beams


74B

A column experiments database was assembled in Section 4.2.6.1 to serve as a


1253H

resourceinthedevelopmentandvalidationoftheDirectStrengthMethodforcolumns
withholes.Inthissection,asimilardatabaseisdevelopedthatsummarizestheelastic
buckling properties and tested strengths of coldformed steel beam experiments with
holes. This database is used in Chapter 8 when developing and verifying DSM for
1254H

beamswithholes.
The beam experiments considered in this study were conducted by Shan,
LaBoube, Schuster, and Batson in the early 1990s and consist of three separate test
sequences(Batson1992;Schuster1992;ShanandLaBoube1994).TestSequences1and2
were performed at the University of MissouriRolla (UMR) and Test Sequence 3 was
executed at the University of Waterloo. Each specimen is made up of two Csections
orientedtoetotoeasdepictedin Figure4.49.3/4x3/4x1/8aluminumanglesconnect
125H

thetopandbottomflangesofthetwochannelswithoneselfdrillingscrewperflange.
Theanglesprovideaclosedbeamsectionthatpreventslateraltorsionalbucklingofthe
individualchannels.

3.1.1 Memberandholedimensionnotation
130B

Beamcrosssectionandholedimensionnotationispresentedin Figure4.50.The
1256H

Csectioninsidecornerradiiareassumedtoequaltwicethemeasuredthicknessofthe
specimen.Twoholeshapeswereconsideredintheexperiments,anindustrystandard

126

slottedholeandatrislottedholewiththecurvedholeendsreplacedbytriangulartips.
Theholesarecenteredinthewebandaremechanicallypunchedat24inchesoncenter
longitudinallywithaholeatthecenterofthespan.

3/4x3/4x1/8 angle(top and bottom)

Cee channel (typ.)

self-drilling screw (typ.)

Figure4.49CrosssectionofbeamspecimenshowingaluminumstrapanglesconnectedtoCflanges

6 in.
B11

B21
D1

Channel 1

H1

Lhole

D21

rhole

Channel 2
hhole

H2

Slotted hole

hhole/2

t
D2

D22

B12

Tri-slotted hole

B22

Figure4.50Csectionandholedimensionnotation

3.1.2 Testedboundaryconditionsandloading
13B

All beam specimens are tested as simply supported in fourpoint bending to


create a region of constant momentbetween the load points at the center of the beam.

127

Thepointloadsareappliedthroughstubchannelsattachedtothebeamwebswithself
drilling screws. The stub channels prevent web crippling by distributing the
concentratedloadandbyrestrainingtheweb.Lateralbracingisprovidedinthevicinity
oftheconstantmomentregionbystrutsconnectingthetopflangealuminumanglestoa
reactionframe.Theendsofthebeamspecimensarelaterallybracedbyverticalrollers.
AsummaryofthebeamtestsetupisprovidedinFigure4.51.
1257H

128


L
6-0

X
P

*
lateral brace point
at angle (typ.)

X
P

a
24 hole spacing (typ.)
12 spacing (typ.) Test Sequence 1, 3
6 spacing (typ.) Test Sequence 2

3/4x3/4x1/8 angle connected with one


screw to each flange (top and bottom)

3/4x3/4x1/8 angle

Channel 2

Channel 1

Section a-a

Figure4.51Experimenttestsetupwithholespacing,locationoflateralbracing,spacingofaluminumanglestraps,andloadpoints

129

3.1.3 FiniteElementModeling
132B

The elastic buckling properties of the 72 beam specimens are obtained with
eigenbucklinganalysesinABAQUS(ABAQUS2004).AllbeamsaremodeledwithS9R5
reduced integration ninenode thin shell elements. Refer to Section 2.1 for a detailed
1258H

discussion of the S9R5 element. Coldformed steel material properties are assumed as
E=29500ksiand=0.3.
Special care is taken to simulate the experimental boundary conditions when
modeling in ABAQUS. The simple supports with vertical roller restraints, the
aluminum angle straps connecting the top and bottom channel flanges, the lateral
bracing of the top flange in the constant moment region, and the application of load
through the webs and are all considered. Figure 4.52 summarizes the ABAQUS
1259H

boundaryconditionassumptions.
Beam end restrained in 2
and 3 (v, w=0)

*
Bottom flange restrained in
1 at support (u=0)

Restrain node at midline of


top flange in 3 (w=0) (Typ.)

Rigid body connection


between top (and bottom)
flange midline nodes (Typ.)

Beam end restrained in 2


and 3 (v, w=0)

Figure4.52Finiteelementmodelboundaryconditionsforbeameigenbucklinganalyses

Tosimulatethesimplysupportedconditionswithverticalrollers,theendsofthe
beams are modeled as warping free except for the bottom flange at one end which is

130

restrainedtopreventlongitudinalrigidbodymovement.Thechannelcrosssectionsare
restrainedfromverticalandlateraltranslationatbothbeamends.
A rigid body restraint is used to model the connectivity between the top and
bottom Csection flanges provided by the aluminum angle straps connected with self
drilling screws. Figure 4.53 demonstrates how each angle is modeled as a rigid body
1260H

made up of one midline flange node from each channel. The rigid body definition
requiresthatthemotion(bothtranslationalandrotational)ofthetwonodesisgoverned
by a single reference node, in this case the midline flange node of Channel 1. The
formulation allows for rigid body motion but requires that the relative position of the
twonodesremainsconstant.Adisadvantageofthisrigidbodyrestraintisthatflange
movementsareonlyrestrainedatthemidlinenodeanddonotsimulatecontactbetween
the channel flange and aluminum angle, which sometimes results in distortional
bucklingmodesthatwouldnotbephysicallypossible.
Rigid body connection
between top (and bottom)
flange midline nodes

Rigid body reference node


1 inch mesh spacing (typ.)

Channel 1

Channel 2

Figure4.53Channelandholemeshingdetailsandmodelingofaluminumanglestraps

ElementmeshingisperformedwithacustombuiltMatlabprogramwrittenbythe
author(SeeAppendixA).Figure4.53providesanexampleofatypicalFEmesh,where
126H

126H

131

theholesaredefinedwithaseriesofelementlinesradiatingfromtheopening. Figure
1263H

4.54 provides a closeup view of the rounded corner meshing of the channels. Two
elementsmodeltheroundedcornersherebecauseS9R5elementshavequadraticshape
functions which allow the initial curved geometry. Refer to Section 2.2 for more
1264H

informationonmodelingroundedcornerswithS9R5elements.
2-ABAQUS S9R5 finite elements used for rounded corners
of channels (max element aspect ratio of 16 to 1)

Figure4.54ABAQUSmeshingdetailsforCsectionroundedcorners

Concentrated loads are applied to the beam specimens through vertical stub
channels connected to the beam webs with selfdrilling screws. To simulate the
distributionoftheloadintoachannelweb,theconcentratedloadisappliedasagroup
ofwebpointloadsinABAQUS. Figure4.55demonstrateshowtheconcentratedloads
1265H

are applied to the beam webs in the finite element models. The web local buckling
restraint(essentiallydoublingupofthewebattheloadingpoint)providedbythestub
channels is not modeled in ABAQUS because it was observed to have a negligible
influenceontheelasticbucklingbehaviorintherelativelylongconstantmomentregions
ofthebeams.

132

Applied load

Stub channel

The transfer of load from the stub channels through the


self-drilling screws is simulated as a series of web point
loads in ABAQUS

Figure4.55ModelingofthebeamconcentratedloadsinABAQUS

3.1.4 Elasticbucklingresultsandmodedefinitions
13B

The beam specimen elastic buckling modes were reviewed in ABAQUS by the
authortoidentifythepurelocal(L)anddistortional(D)bucklingmodesaswellasany
mixedelasticmodescreatedbytheadditionofwebholes.Lateraltorsionalbucklingis
restrainedbythetopflangelateralbracingandaluminumanglestraps(seeFigure4.51),
126H

althoughotherpossibleglobal(G)bucklingmodesarepossibleasdiscussedinSection
4.3.1.4.3.
1267H

The mode identification process for beams with holes is guided by the
experiences obtained in Section 4.2.4 for coldformed steel compression members with
1268H

holes. Csection columns with web holes exhibited unique mixed buckling modes
wheredistortionoftheflangesneartheholemixeswithlocalbuckling(LHmode).In

133

thisbeamstudymixedlocaldistortionalmodesareagainobserved,aswellaslocalweb
holemodesinitiatedbythecompressioncomponentofthestressgradientfrombending.

4.3.1.4.1 LocalBuckling
195B

Slotted holes in the beam specimen webs initiate unique local buckling modes
and reduce the critical elastic local buckling moment Mcrl in most cases. The shallow
beamspecimenwithoutholes(nominaldepthof2.5inches)in Figure4.56exhibitslocal
1269H

buckling in both the top flange and web. The addition of slotted web holes creates a
newlocalbucklingmode,theLH2mode.TheLH2modeoccurswhenthestripofweb
abovetheholebucklesintwohalfwaves.Thismodeoccursbecausethefundamental
local buckling halfwavelength of the crosssection, Lcrl, is less than the length of the
hole.ThecriticalelasticbucklingmomentfortheLH2modeis8percentlessthanthatof
the pure L mode, suggesting that this local hole mode may influence the load
deformationresponseofthebeam.
Figure 4.57 compares the elastic buckling behavior of a slightly deeper beam
1270H

(nominal depth of 3.625 inches) with and without holes. The addition of slotted web
holes again creates the LH2 mode with a critical elastic buckling moment that is 17
percentlessthanthepureLmode. Figure4.58, Figure4.59,and Figure4.60summarize
127H

127H

1273H

the influence of slotted holes on the local buckling behavior of deeper beams with
nominal heights of 6 inches, 8 inches, and 12 inches respectively. The LH mode is
identifiedinthesedeeperbeamdepthsasthebucklingofthestripofwebabovethehole
intoasinglehalfwave.TheLH2modeisobservedinthe6in.and8in.deepbeamsbut

134

withahighercriticalelasticbucklingmomentthattheLHmode.TheLH2modeisnot
observedforthe12in.deepspecimensinceLcrlforthisspecimenexceedsthelengthof
thehole.

LH2
Mcrl/Myg=0.77

L
Mcrl/Myg=0.82

L
Mcrl/M yg=0.83

Figure4.56Localbucklingmodesforspecimen2B,20,1&2(H)withandwithoutholes

LH2
Mcrl/Myg=2.49

L
Mcrl/Myg=3.00

L
Mcrl/Myg=3.02

Figure4.57Localbucklingmodesforspecimen3B,14,1&2(H)withandwithoutholes

135

LH
Mcrl/Myg=0.75

L
Mcrl/Myg=1.05
LH2
Mcrl/Myg=0.87

L
Mcrl/Myg=1.07

Figure4.58Localbucklingmodesforspecimen6B,18,1&2(H)withandwithoutholes

LH
Mcrl/Myg=0.63

L
Mcrl/Myg=0.78
LH2
Mcrl/Myg=0.72

L
Mcrl/Myg=0.79

Figure4.59LocalbucklingmodesforspecimenBP40(H)withandwithoutholes

136

LH
Mcrl/Myg=0.85

L
Mcrl/Myg=0.96

L
Mcrl/Myg=0.89

Figure4.60Localbucklingmodesforspecimen12B,16,1&2(H)withandwithoutholes

4.3.1.4.2 DistortionalBuckling
196B

Figure 4.61 compares the influence of slotted web holes on the distortional
1274H

bucklingofashallowbeamspecimen(nominalheightof2.5inches).Forthespecimen
withholes,auniqueDH+Lmodeisobservedwithacriticalelasticbucklingmoment20
percentlessthatthepureDmode.ThismodehassimilarcharacteristicstotheLHmode
in beams (see Section 4.3.1.4.1), especially the buckling of the strip above the hole into
1275H

one halfwave. The DH+L mode is expressed more as a distortional mode though
because the compression flange is wide relative to the unstiffened strip. The D mode
without holes becomes a mixed distortionallocal mode (D+L) when holes are added,
although the critical elastic buckling moment is not significantly affected in this case.
This specimen is sensitive to mixing of local and distortional modes because of the

137

relativelythinsteelsheetthicknesstof0.0346inches.Itisalsonotedthattheholehas
onlyasmallinfluenceonthepureDmodehalfwavelengthforthisspecimen.
The DH distortional buckling mode at the hole is also observed for a slightly
deeper beam specimen (nominal height of 3.625 inches) in Figure 4.62. The sheet
1276H

thicknessforthesechannelsisroughlydoublethatofthepreviouslydiscussedspecimen
(t=0.71inches)andtheholedepthisunchanged.McrlishigherthanMcrdbecauseofthe
increased thickness, resulting in DH and D modes without local buckling interaction
whentheslottedholesarepresent.ThecriticalelasticbucklingmomentoftheDHmode
is13percentlessthanthatofthepureDmode.

DH+L
Mcrd/Myg=0.83
D
Mcrd/Myg=1.08
Half wavelength=13 inches

D+L
Mcrd/Myg=1.03
Half wavelength=14 inches

Figure4.61Distortionalbucklingmodesforspecimen2B,20,1&2(H)withandwithoutholes

138

D
Mcrd/Myg=2.31
Half wavelength=12 inches

DH
Mcrd/Myg=2.00

D
Mcrd/Myg=2.31
Half wavelength=12 inches

Figure4.62Distortionalbucklingmodesforspecimen3B,14,1&2(H)withandwithoutholes

Figure4.63andFigure4.64comparetheinfluenceofslottedwebholesonbeams
127H

1278H

with nominal heights of six inches and eight inches respectively, both having a steel
sheet thickness of t=0.047 inches. For these specimens the unstiffened strip buckling
modeabovetheholeisidentifiedasLHbuckling(seeFigure4.58,Figure4.59)insteadof
1279H

1280H

DHbucklingbecausethemajorityofthebucklingdeformationoccursintheweb.The
similaritiesbetweentheLHandDHmodescanmakethemdifficulttoclassifyinsome
cases.Researchonamechanicsbasedmodalidentificationmethodisunderway.

D
Mcrd/Myg=1.56
Half wavelength=12 inches

D
Mcrd/Myg=1.56
Half wavelength=12 inches

Figure4.63Distortionalbucklingmodesforspecimen6B,18,1&2(H)withandwithoutholes

139

D
Mcrd/Myg=1.02
Half wavelength=12 inches

D
Mcrd/Myg=1.00
Half wavelength=12 inches

Figure4.64DistortionalbucklingmodesforspecimenBP540(H)withandwithoutholes

D+L
Mcrd/Myg=1.03
Half wavelength=11 inches

D
Mcrd/Myg=1.02
Half wavelength=12 inches

Figure4.65Distortionalbucklingmodesforspecimen12B,16,1&2(H)withandwithoutholes

Thedistortionalbucklingmodesofthedeepestbeamspecimenconsideredinthis
study (nominal depth of 12 inches) are provided in Figure 4.65. Identifying the
128H

distortional buckling modes for the channels making up this beam are inherently
challenging because even for a member without holes, there is not a clear distinction
betweentheLandDmodes.ThecriticalelasticbucklingmomentsforasingleCsection
fromthebeamcrosssectionareprovidedatvarioushalfwavelengthsfromafinitestrip
analysis (CUFSM) in Figure 4.66 (including the modal participation factors calculated
128H

withtheconstrainedfinitestripmethod).Onlyoneminimumexists,suggestingthatthe
modesatorneartheminimumbucklingloadareamixtureofLandDmodes.Themost
suitablemodeidentifiedbytheauthor(forthespecimenwithoutahole)in Figure4.65
1283H

alternatesbetweenlargerdistortionalhalfwavesandshorterlocalbucklinghalfwaves
intheconstantmomentregionofthechannels.Forthespecimenwiththewebholesin

140

Figure 4.65, the local halfwaves are not present and the mode resembles more of a
1284H

pure D mode. The DH mode is not observed for this specimen, which is consistent
with the buckling behavior of stiffened elements in bending (see Figure 3.25).
1285H

Unstiffenedstripbuckling(theplatemodethatishypothesizedtoinitiatetheDHmode
inbeams)doesnotoccurwhenhhole/hissmall.

D
This plot summarizes the
modal participation (L, D,
G, O) as a function of
half-wavelength

Figure4.66Elasticbucklingcurvefor12deepspecimenwithmodalparticipationsummarized,notethat
selectedLandDaremixedlocaldistortionalmodes

4.3.1.4.3 Globalbuckling
197B

Lateraltorsional buckling is a common global (G) elastic buckling mode in


beams, although this mode is eliminated for the specimens considered here by
connectingthetwoCsectionstoetotoewithaluminumanglesandbyprovidinglateral
bracing at the compression flange in the constant moment region of the beams (see
Figure 4.51). Twisting of an individual channel about its longitudinal axis is still
1286H

possible though, even with the top flange restrained. Figure 4.67 depicts the potential
1287H

twistingmode.CUFSMisusedtoconservativelyquantifytheelasticbucklingmoment
forthismode,anditisdeterminedthatMcreismorethantentimestheyieldmomentMy

141

forthespecimensinthisstudy.SincetheMcredoesnotinfluencetheDSMpredictionas
longasMcre>2.78My,theglobaltwistingmodeisnotsummarizedinthedatabase.
Lateral bracing of top flange

Figure4.67Possibleglobalbucklingmodeoccursaboutthecompressionflangelateralbracepoint

3.1.5 Elasticbucklingdatabaseforbeamswithholes
134B

Table 4.11 summarizes the dimensions and material properties of each channel
128H

making up the beam (Channel 1 and Channel 2), including cross section and hole
dimensions,testedultimatepointloadPtest(foreachchannel)andultimatemomentMtest
(for each channel), tested specimen yield stress Fy, specimen yield moment My,g
(calculated with the gross crosssection), and My,net (calculated with the net cross
section). Fy varies from 22.0 ksi to 93.3 ksi with a mean of 48.6 ksi and standard
deviationof14ksi.Thislargevariationinyieldstresswassomewhatunexpected.

ABAQUS eigenbuckling results are summarized in Table 4.12 for each channel
1289H

considering the experiment boundary conditions both with and without holes. These
resultsareusedinSection 4.3.1.6toevaluatetheinfluenceofholesonMcrlandMcrd.The
1290H

CUFSM elastic buckling results are also provided, including the fundamental
distortional halfwavelength Lcrd, which are used in Section 4.3.1.7 to evaluate the
129H

influenceofexperimentboundaryconditionsonMcrlandMcrdandthedistortionalhalf
wavelength.

142

Table 4.10 presents the crosssection parameter ranges of the beam Csections
129H

contained in the experiment database. All of the beam specimens have crosssection
dimensions that meet the DSM prequalification standards for ultimate strength
prediction summarized in Table 4.9 (AISIS100 2007). Four of the beam specimens
1293H

exceedtheyieldstressprequalificationlimitofFy<70ksi.

Table4.9DSMprequalificationlimitsforbeamCsections

DSM
prequalification limit

Beam parameter
Web slenderness

H/t<321

Flange slenderness B/t<75


Lip slenderness

0<D/t<34

Web / flange

1.5<H/B<17

Lip / flange
Yield stress

0<D/B<0.70
Fy<70 ksi

Table4.10Parameterrangesforbeamspecimenswithholes

min
max

D/t
5.5
22.1

H/t
40.5
257.1

B/t
16.3
58.3

H/B
1.5
7.7

143

D/B
0.18
0.42

hhole/h
0.13
0.67

F y (ksi)
22.0
93.3

Table4.11Beamexperimentcrosssectiondimensions,materialproperties,andtestedstrengths
Study and Specimen Name
Test
Sequence
Shan and LaBoube 1994

Schuster 1992

1
1
1
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
3
3

Member Length, Load


Location and Thickness
L

2,16,1&2(H)
2,20,1&2(H)
2,20,3,4(H)
3,14,1&2(H)
3,14,3&4(H)
3,18,1&2(H)
3,18,3&4(H)
3,20,1&2(H)
3,20,3&4(H)
12,14,1&2(H)
12,14,3&4(H)
12,16,1&2(H)
12,16,3&4(H)
2B,16,1&2(H)
2B,16,3&4(H)
2B,20,1&2(H)
2B,20,3&4(H)
3B,14,1&2(H)
3B,14,3&4(H)
3B,18,1&2(H)
3B,18,3&4(H)
3B,20,1&2(H)
3B,20,3&4(H)
3B,20,5&6(H)
3B,20,1&2(T)
3B,20,3&4(T)
6B,18,1&2(H)
6B,18,3&4(H)
6C,18,1&2(H)
6C,18,3&4(H)
6D,18,1&2(H)
6D,18,3&4(H)
6B,20,1&2(H)
8A,14,1&2(H)
8A,14,3&4(H)
8A,14,5&6(H)
8A,14,7&8(H)
8A,14,9&10(H)
8B,14,1&2(T)
8B,14,3&4(T)
8B,14,5&6(T)
8B,14,7&8(T)
8D,14,1&2(T)
8D,14,3&4(T)
8B,18,1&2(H)
8D,18,1&2(H)
8D,18,3&4(H)
8A,20,1&2(H)
8A,20,3&4(H)
8B,20,1&2(T)
8B,20,3&4(T)
8B,20,5&6(T)
8B,20,7&8(T)
8D,20,1&2(T)
8D,20,3&4(T)
8D,20,5&6(T)
12B,16,1&2(H)
12B,16,3&4(H)
12B,16,5&6(H)
12B,16,7&8(H)
BP440(H)
BP540(H)
BP640(H)
BP765(H)
BP865(H)
BP965(H)
CP440(T)
CP540(T)
CP640(T)
CP765(T)
CP865(T)
CP965(T)

in.
150.0
150.0
150.0
150.0
150.0
150.0
150.0
150.0
150.0
192.0
192.0
192.0
192.0
150.0
150.0
150.0
150.0
150.0
150.0
150.0
150.0
150.0
150.0
150.0
150.0
150.0
192.0
192.0
192.0
192.0
192.0
192.0
192.0
192.0
192.0
192.0
192.0
192.0
192.0
192.0
192.0
192.0
192.0
192.0
192.0
192.0
192.0
192.0
192.0
192.0
192.0
192.0
192.0
192.0
192.0
192.0
192.0
192.0
192.0
192.0
168.0
168.0
168.0
168.0
168.0
168.0
168.0
168.0
168.0
168.0
168.0
168.0

shear
span
in.
39.0
39.0
39.0
39.0
39.0
39.0
39.0
39.0
39.0
60.0
60.0
60.0
60.0
39.0
39.0
39.0
39.0
39.0
39.0
39.0
39.0
39.0
39.0
39.0
39.0
39.0
60.0
60.0
60.0
60.0
60.0
60.0
60.0
60.0
60.0
60.0
60.0
60.0
60.0
60.0
60.0
60.0
60.0
60.0
60.0
60.0
60.0
60.0
60.0
60.0
60.0
60.0
60.0
60.0
60.0
60.0
60.0
60.0
60.0
60.0
48.0
48.0
48.0
48.0
48.0
48.0
48.0
48.0
48.0
48.0
48.0
48.0

t
in.
0.062
0.039
0.039
0.077
0.077
0.044
0.044
0.044
0.044
0.098
0.098
0.055
0.055
0.059
0.059
0.033
0.033
0.071
0.071
0.044
0.044
0.036
0.036
0.036
0.029
0.029
0.046
0.046
0.048
0.048
0.046
0.046
0.033
0.074
0.074
0.074
0.065
0.065
0.067
0.067
0.065
0.065
0.065
0.065
0.045
0.046
0.046
0.031
0.031
0.031
0.031
0.031
0.031
0.043
0.043
0.043
0.060
0.060
0.060
0.060
0.047
0.047
0.047
0.047
0.047
0.047
0.048
0.048
0.048
0.048
0.048
0.048

Yield Stress and Moment of Inertia

Material
Properties
E
ksi
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500
29500

nu

0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3

Hole Dimensions

Hole Type

Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Tri-slotted
Tri-slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Tri-slotted
Tri-slotted
Tri-slotted
Tri-slotted
Tri-slotted
Tri-slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Tri-slotted
Tri-slotted
Tri-slotted
Tri-slotted
Tri-slotted
Tri-slotted
Tri-slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Tri-slotted
Tri-slotted
Tri-slotted
Tri-slotted
Tri-slotted
Tri-slotted

L hole
in.
2.000
2.000
2.000
4.000
4.000
4.000
4.000
4.000
4.000
4.000
4.000
4.000
4.000
4.000
4.000
4.000
4.000
4.000
4.000
4.000
4.000
4.000
4.000
4.000
4.500
4.500
4.000
4.000
4.000
4.000
4.000
4.000
4.000
4.000
4.000
4.000
4.000
4.000
4.500
4.500
4.500
4.500
4.500
4.500
4.000
4.000
4.000
4.000
4.000
4.500
4.500
4.500
4.500
4.500
4.500
4.500
4.000
4.000
4.000
4.000
4.02
4.02
4.02
4.53
4.53
4.53
4.65
4.65
4.65
4.61
4.61
4.61

h hole
in.
0.750
0.750
0.750
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.500
1.5
1.5
1.5
2.48
2.48
2.48
1.69
1.69
1.69
2.52
2.52
2.52

Cross Section Dimensions

r hole
in.
0.375
0.375
0.375
0.750
0.750
0.750
0.750
0.750
0.750
0.750
0.750
0.750
0.750
0.750
0.750
0.750
0.750
0.750
0.750
0.750
0.750
0.750
0.750
0.750
----0.750
0.750
0.750
0.750
0.750
0.750
0.750
0.750
0.750
0.750
0.750
0.750
------------0.750
0.750
0.750
0.750
0.750
--------------0.750
0.750
0.750
0.750
0.750
0.750
0.750
0.750
0.750
0.750
-------------

H1
in.
2.510
2.500
2.510
3.680
3.690
3.750
3.650
3.650
3.670
12.080
12.050
11.960
12.070
2.460
2.470
2.420
2.420
3.650
3.640
3.610
3.620
3.610
3.610
3.600
3.560
3.560
6.060
6.050
5.960
5.950
6.020
6.020
5.920
8.060
8.070
8.070
8.030
8.040
8.050
8.050
8.020
8.030
7.950
7.950
7.950
8.000
8.000
7.930
7.930
7.970
7.960
7.950
7.950
7.940
7.940
7.950
11.950
11.980
11.960
11.970
7.99
7.99
7.99
7.99
7.99
7.99
7.99
7.99
8.03
7.99
8.03
7.99

144

H2
in.
2.510
2.480
2.520
3.680
3.690
3.650
3.640
3.710
3.690
12.070
12.000
11.970
11.960
2.460
2.460
2.420
2.430
3.620
3.630
3.630
3.630
3.600
3.610
3.600
3.570
3.560
6.050
6.020
5.960
5.980
6.020
6.020
5.920
8.060
8.070
8.070
8.030
8.040
8.050
8.040
8.020
8.030
7.960
7.950
7.940
8.000
8.000
7.930
7.920
7.970
7.960
7.950
7.950
7.940
7.940
7.950
11.950
12.020
11.970
11.960
7.99
7.99
7.99
7.99
7.99
7.99
7.99
7.99
8.03
7.99
7.99
7.99

Channel
1
2

B 11

B 21

B 12

B 22

D 11

D 21

D 12

D 22

in.
1.610
1.600
1.590
1.650
1.630
1.560
1.560
1.560
1.560
1.640
1.640
1.570
1.560
1.620
1.630
1.630
1.630
1.620
1.630
1.610
1.620
1.630
1.640
1.630
1.620
1.620
1.620
1.620
1.980
1.970
2.420
2.430
1.630
1.380
1.380
1.370
1.390
1.390
1.640
1.640
1.630
1.630
2.480
2.470
1.590
2.420
2.420
1.380
1.370
1.630
1.630
1.630
1.630
2.490
2.460
2.490
1.630
1.630
1.630
1.630
1.61
1.61
1.61
1.58
1.61
1.61
1.58
1.58
1.61
1.61
1.58
1.61

in.
1.610
1.600
1.620
1.640
1.620
1.560
1.580
1.640
1.590
1.630
1.600
1.570
1.570
1.630
1.620
1.640
1.640
1.660
1.620
1.650
1.660
1.620
1.630
1.630
1.650
1.680
1.620
1.620
1.990
1.980
2.430
2.430
1.620
1.380
1.380
1.380
1.390
1.380
1.630
1.640
1.640
1.630
2.500
2.490
1.580
2.450
2.450
1.390
1.380
1.640
1.630
1.630
1.630
2.450
2.460
2.460
1.630
1.630
1.630
1.630
1.61
1.61
1.61
1.58
1.58
1.58
1.58
1.61
1.61
1.61
1.61
1.61

in.
1.630
1.600
1.580
1.630
1.640
1.570
1.560
1.550
1.550
1.690
1.670
1.570
1.570
1.620
1.620
1.630
1.630
1.630
1.620
1.650
1.650
1.630
1.640
1.620
1.680
1.690
1.550
1.550
1.980
1.990
2.430
2.430
1.520
1.380
1.380
1.380
1.390
1.380
1.640
1.640
1.640
1.630
2.470
2.470
1.580
2.440
2.450
1.380
1.390
1.630
1.620
1.630
1.640
2.450
2.440
2.450
1.630
1.620
1.630
1.620
1.61
1.61
1.61
1.58
1.61
1.58
1.58
1.58
1.58
1.61
1.58
1.63

in.
1.610
1.600
1.600
1.630
1.630
1.580
1.570
1.590
1.610
1.630
1.710
1.560
1.580
1.610
1.630
1.620
1.620
1.630
1.630
1.620
1.640
1.620
1.630
1.630
1.600
1.610
1.550
1.550
1.990
1.980
2.430
2.430
1.530
1.380
1.380
1.370
1.400
1.380
1.640
1.640
1.630
1.630
2.490
2.480
1.580
2.430
2.430
1.380
1.370
1.620
1.630
1.630
1.630
2.490
2.480
2.480
1.630
1.630
1.630
1.630
1.61
1.61
1.61
1.58
1.58
1.58
1.58
1.58
1.58
1.61
1.61
1.61

in.
0.400
0.420
0.360
0.570
0.530
0.580
0.560
0.520
0.600
0.690
0.650
0.500
0.420
0.470
0.470
0.420
0.420
0.540
0.540
0.510
0.500
0.460
0.460
0.460
0.590
0.580
0.470
0.470
0.640
0.600
0.700
0.700
0.440
0.490
0.500
0.410
0.430
0.460
0.630
0.640
0.640
0.660
0.640
0.660
0.470
0.610
0.600
0.410
0.450
0.610
0.620
0.610
0.610
0.640
0.640
0.620
0.530
0.470
0.510
0.480
0.47
0.47
0.47
0.47
0.47
0.47
0.51
0.51
0.53
0.51
0.51
0.53

in.
0.450
0.410
0.420
0.550
0.530
0.560
0.570
0.560
0.560
0.600
0.640
0.610
0.530
0.460
0.520
0.420
0.410
0.550
0.470
0.520
0.500
0.470
0.470
0.460
0.640
0.630
0.470
0.480
0.590
0.650
0.620
0.700
0.470
0.480
0.410
0.500
0.480
0.440
0.640
0.640
0.630
0.610
0.480
0.480
0.470
0.690
0.700
0.440
0.430
0.610
0.580
0.600
0.620
0.590
0.590
0.620
0.540
0.500
0.500
0.550
0.47
0.47
0.47
0.47
0.47
0.47
0.51
0.51
0.51
0.51
0.51
0.51

in.
0.420
0.420
0.470
0.560
0.620
0.580
0.540
0.550
0.520
0.600
0.650
0.520
0.580
0.510
0.520
0.500
0.500
0.490
0.490
0.500
0.520
0.460
0.470
0.460
0.620
0.620
0.500
0.500
0.590
0.640
0.620
0.610
0.440
0.410
0.410
0.490
0.480
0.450
0.670
0.660
0.640
0.610
0.470
0.450
0.480
0.690
0.700
0.450
0.430
0.600
0.580
0.600
0.610
0.590
0.590
0.620
0.520
0.550
0.510
0.560
0.47
0.47
0.47
0.47
0.47
0.47
0.51
0.51
0.51
0.51
0.51
0.51

in.
0.430
0.410
0.410
0.520
0.550
0.540
0.540
0.560
0.590
0.620
0.640
0.430
0.530
0.510
0.460
0.500
0.500
0.500
0.540
0.500
0.520
0.470
0.470
0.470
0.610
0.570
0.500
0.510
0.640
0.630
0.700
0.620
0.420
0.430
0.500
0.410
0.450
0.480
0.660
0.650
0.630
0.660
0.620
0.610
0.470
0.620
0.600
0.430
0.440
0.620
0.620
0.610
0.620
0.640
0.650
0.630
0.530
0.530
0.520
0.490
0.47
0.47
0.47
0.47
0.47
0.47
0.51
0.51
0.51
0.51
0.53
0.51

in.
0.124
0.078
0.078
0.154
0.154
0.088
0.088
0.088
0.088
0.196
0.196
0.110
0.110
0.118
0.118
0.066
0.066
0.142
0.142
0.088
0.088
0.072
0.072
0.072
0.058
0.058
0.092
0.092
0.096
0.096
0.092
0.092
0.066
0.148
0.148
0.148
0.130
0.130
0.134
0.134
0.130
0.130
0.130
0.130
0.090
0.092
0.092
0.062
0.062
0.062
0.062
0.062
0.062
0.086
0.086
0.086
0.120
0.120
0.120
0.120
0.094
0.094
0.094
0.094
0.094
0.094
0.096
0.096
0.096
0.096
0.096
0.096

Fy
ksi
37.23
33.70
33.70
63.72
63.72
46.92
46.92
46.82
46.82
35.93
35.93
49.11
49.11
53.59
53.59
67.15
67.15
81.36
81.36
53.13
53.13
63.71
63.71
63.71
25.51
25.51
47.17
47.17
75.08
75.08
30.77
30.77
93.26
31.04
31.04
31.04
56.29
56.29
32.58
32.58
53.14
53.14
54.71
54.71
72.32
22.00
22.00
37.96
37.96
44.89
44.89
44.89
44.89
38.59
38.59
38.59
60.64
60.64
60.64
60.64
38.87
38.87
38.87
38.87
38.87
49.02
49.02
49.02
49.02
49.02
49.02
49.02

Channel
1
2

Experimental
Results

M y,g

M y,g

M y,net

M y,net

P test

M test

k*in
12.25
7.13
7.17
43.11
43.83
19.05
18.18
18.11
18.14
167.26
167.02
124.61
126.99
16.68
16.80
11.99
11.99
49.80
49.49
20.81
20.98
20.36
20.48
20.22
6.82
6.84
37.34
37.26
70.77
71.28
32.19
32.16
50.95
53.04
53.16
53.77
86.17
85.70
58.37
58.29
91.54
91.07
113.41
113.04
82.33
33.87
33.97
27.95
27.94
37.40
37.13
37.27
37.42
54.64
54.46
54.95
169.78
170.43
169.55
170.58
46.76
46.76
46.76
46.33
46.76
58.55
60.19
60.19
60.81
60.75
60.62
61.09

k*in
12.18
7.04
7.19
42.72
43.07
18.30
18.20
18.90
18.93
164.89
166.45
123.30
125.25
16.62
16.57
11.95
12.01
49.51
49.91
20.81
21.05
20.25
20.40
20.33
6.68
6.64
37.26
37.10
71.51
71.59
32.55
32.20
50.90
53.24
53.93
53.02
86.08
86.09
58.25
58.09
91.20
91.71
116.36
115.71
81.96
33.58
33.48
27.90
27.80
37.47
37.42
37.32
37.41
55.30
55.30
55.27
170.15
171.33
169.96
169.32
46.76
46.76
46.76
46.33
46.33
58.43
60.19
60.32
60.75
60.75
60.98
60.75

k*in
12.18
7.09
7.14
42.35
43.05
18.74
17.86
17.79
17.83
167.09
166.85
124.48
126.86
15.93
16.06
11.46
11.46
48.89
48.58
20.44
20.61
20.00
20.12
19.86
6.70
6.73
37.14
37.06
70.43
70.94
32.06
32.02
50.66
52.88
53.00
53.61
85.91
85.45
58.22
58.14
91.30
90.83
113.16
112.79
82.10
33.80
33.90
27.86
27.85
37.30
37.03
37.17
37.32
54.52
54.34
54.83
169.61
170.26
169.38
170.41
46.64
46.64
46.64
45.74
46.18
57.82
59.95
59.95
60.57
59.96
59.84
60.30

k*in
12.11
7.00
7.16
41.96
42.30
17.98
17.88
18.59
18.61
164.73
166.28
123.17
125.12
15.88
15.84
11.43
11.49
48.61
48.99
20.45
20.68
19.89
20.04
19.97
6.56
6.52
37.06
36.89
71.17
71.25
32.41
32.07
50.61
53.08
53.77
52.86
85.83
85.83
58.10
57.93
90.95
91.47
116.10
115.45
81.73
33.51
33.41
27.81
27.71
37.37
37.32
37.22
37.31
55.18
55.18
55.15
169.98
171.16
169.79
169.15
46.64
46.64
46.64
45.74
45.74
57.69
59.95
60.08
60.52
59.96
60.19
59.96

kips
1.04
0.46
0.46
3.7
3.54
1.35
1.37
1.35
1.43
7.16
7.50
4.38
4.79
1.345
1.36
0.6
0.635
4.31
4.255
1.6
1.51
1.2
1.1
1.335
0.425
0.455
1.64
1.7
3.425
3.445
1.67
1.7
1.15
3.675
3.7
3.64
4.37
4.31
3.225
3.89
3.735
5.375
5.895
5.925
2.76
2.1
1.84
1.005
0.985
1.37
1.4
1.4
1.4
2.5
2.7
2.6
6.5
6.4
6.4
6.7
3.2
3.1
3.2
3.1
3.2
3.2
3.5
3.3
3.5
3.4
3.4
3.4

k*in
10.1
4.5
4.5
36.1
34.5
13.2
13.4
13.2
13.9
107.4
112.5
65.7
71.9
13.1
13.3
5.9
6.2
42.0
41.5
15.6
14.7
11.7
10.7
13.0
4.1
4.4
24.6
25.5
51.4
51.7
25.1
25.5
17.3
55.1
55.5
54.6
65.6
64.7
48.4
58.4
56.0
80.6
88.4
88.9
41.4
31.5
27.6
15.1
14.8
20.6
20.6
21.4
20.7
38.1
39.8
39.0
97.3
96.6
95.8
100.0
37.9
36.8
38.2
37.7
38.2
38.2
41.4
39.4
41.6
41.3
40.9
40.8

Table4.12Beamexperimentelasticbucklingresults

Channel 1
M cr l ,LH M cr l ,LH2 M cr l ,L

Test Sequence

Shan and LaBoube 1994

Schuster 1992

1
1
1
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
3
3

ABAQUS elastic buckling WITHOUT holes,


experiment boundary conditions

ABAQUS elastic buckling with holes, experiment boundary conditions

Study and Specimen Name

2,16,1&2(H)
2,20,1&2(H)
2,20,3,4(H)
3,14,1&2(H)
3,14,3&4(H)
3,18,1&2(H)
3,18,3&4(H)
3,20,1&2(H)
3,20,3&4(H)
12,14,1&2(H)
12,14,3&4(H)
12,16,1&2(H)
12,16,3&4(H)
2B,16,1&2(H)
2B,16,3&4(H)
2B,20,1&2(H)
2B,20,3&4(H)
3B,14,1&2(H)
3B,14,3&4(H)
3B,18,1&2(H)
3B,18,3&4(H)
3B,20,1&2(H)
3B,20,3&4(H)
3B,20,5&6(H)
3B,20,1&2(T)
3B,20,3&4(T)
6B,18,1&2(H)
6B,18,3&4(H)
6C,18,1&2(H)
6C,18,3&4(H)
6D,18,1&2(H)
6D,18,3&4(H)
6B,20,1&2(H)
8A,14,1&2(H)
8A,14,3&4(H)
8A,14,5&6(H)
8A,14,7&8(H)
8A,14,9&10(H)
8B,14,1&2(T)
8B,14,3&4(T)
8B,14,5&6(T)
8B,14,7&8(T)
8D,14,1&2(T)
8D,14,3&4(T)
8B,18,1&2(H)
8D,18,1&2(H)
8D,18,3&4(H)
8A,20,1&2(H)
8A,20,3&4(H)
8B,20,1&2(T)
8B,20,3&4(T)
8B,20,5&6(T)
8B,20,7&8(T)
8D,20,1&2(T)
8D,20,3&4(T)
8D,20,5&6(T)
12B,16,1&2(H)
12B,16,3&4(H)
12B,16,5&6(H)
12B,16,7&8(H)
BP440(H)
BP540(H)
BP640(H)
BP765(H)
BP865(H)
BP965(H)
CP440(T)
CP540(T)
CP640(T)
CP765(T)
CP865(T)
CP965(T)

k*in
65.6
17.0
16.9
------------------------9.3
9.3
----28.3
28.1
15.6
15.6
15.6
8.6
8.6
28.0
28.1
36.4
35.5
35.0
34.9
10.7
103.0
104.6
98.0
69.2
70.8
86.1
86.5
79.2
80.1
96.1
96.7
25.9
34.0
33.9
8.0
8.1
8.9
9.0
8.9
8.9
28.3
28.1
28.1
59.2
57.6
59.0
57.9
29.3
29.3
29.3
26.9
27.4
27.4
30.2
30.3
31.0
30.7
30.4
31.1

k*in
------153.7
149.2
29.6
29.7
29.4
29.8
--------53.8
53.4
9.3
9.3
124.1
--29.5
29.3
16.2
16.1
16.1
8.7
8.7
32.6
32.6
41.9
41.1
39.7
39.6
12.3
117.5
117.9
109.1
81.0
80.2
98.7
99.2
90.7
91.6
109.5
110.0
29.4
38.3
38.2
9.1
9.2
10.1
10.2
10.1
10.1
32.1
31.9
31.9
--------33.5
33.5
33.5
30.8
30.8
31.2
34.9
35.0
35.6
35.2
34.8
35.4

k*in
67.0
17.0
17.1
193.1
194.3
38.6
38.2
37.7
38.3
248.8
241.7
44.6
41.9
57.3
58.5
9.9
9.7
150.2
151.0
36.7
36.4
20.2
20.1
20.1
10.4
10.5
39.9
39.9
50.6
49.8
47.3
45.3
15.1
131.3
134.2
128.6
88.5
90.3
110.6
111.1
101.7
102.6
121.2
121.9
32.8
42.3
42.2
10.1
10.3
11.4
11.4
11.4
11.3
35.6
35.4
35.3
62.0
60.4
61.6
60.7
37.3
37.3
37.3
37.1
37.4
37.4
39.9
39.9
40.5
40.3
39.8
40.5

M crd,DH
k*in
50.3
----115.7
109.9
28.3
28.5
28.0
29.2
--------43.5
43.0
10.0
10.2
99.5
102.0
35.7
35.3
19.7
19.6
19.6
--------59.8
57.3
47.3
47.4
--133.0
133.0
120.3
90.9
90.6
------------------------------53.6
56.5
56.6
---------------------------------

Channel 1

Channel 2
M crd,D
k*in
38.5
15.1
13.0
106.5
101.8
39.3
36.6
33.8
40.3
250.9
243.5
51.0
44.0
39.9
41.6
12.4
12.3
114.9
114.3
38.9
40.5
24.4
23.0
23.0
15.9
16.1
58.3
58.6
73.2
74.9
59.2
66.7
28.1
129.6
130.2
119.9
86.8
89.0
152.3
153.3
137.1
140.3
137.4
138.3
46.7
76.7
75.9
16.0
16.7
30.4
28.6
27.9
28.3
62.6
62.6
63.3
71.6
69.0
75.6
69.6
47.2
47.2
47.2
47.0
47.6
47.7
51.7
52.6
54.8
53.5
52.8
53.6

L crd *
in.
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
14.0
14.0
14.0
14.0
12.0
12.0
12.0
14.0
14.0
16.0
16.0
25.0
25.0
12.0
12.0
20.0
23.0
24.0
27.0
12.0
6.0
6.0
5.0
5.0
6.0
12.0
12.0
12.0
14.0
24.0
24.0
10.0
27.0
24.0
12.0
10.0
17.0
14.0
14.0
13.0
24.0
24.0
24.0
11.0
11.0
11.0
11.0
12.0
12.0
12.0
12.0
12.0
12.0
11.0
12.0
12.0
12.0
12.0
12.0

M cr l ,LH M cr l ,LH2 M cr l ,L
k*in
66.3
16.8
17.0
------------------------9.3
9.3
----28.4
28.1
15.6
15.6
15.6
8.7
--28.0
28.1
35.7
36.3
34.0
34.9
10.8
103.0
98.3
104.7
71.5
69.5
86.1
86.5
79.2
78.3
90.2
90.2
25.8
35.2
35.3
8.1
8.1
8.9
8.8
8.9
8.9
27.5
27.6
27.8
59.2
58.5
58.4
59.9
29.3
29.3
29.3
26.9
27.0
27.0
30.2
30.7
30.7
30.7
30.8
30.7

k*in
------153.7
149.2
29.6
29.8
30.2
29.8
--------53.2
54.6
9.3
9.3
124.1
--29.5
29.3
16.2
16.1
16.1
8.7
8.7
32.6
32.6
41.2
41.9
38.8
39.6
12.4
116.6
109.6
118.0
78.5
78.7
98.7
99.2
90.7
89.7
103.9
103.9
29.3
39.5
39.6
9.1
9.1
10.2
10.0
10.1
10.1
31.3
31.4
31.6
--------33.5
33.5
33.5
30.8
30.8
30.8
34.9
35.3
35.3
35.2
35.2
35.2

k*in
67.0
17.0
17.2
193.1
194.2
38.6
38.2
37.6
38.4
234.0
236.6
46.8
45.0
57.3
58.5
9.9
9.7
149.3
150.1
36.7
36.4
20.2
20.1
20.1
10.4
10.2
39.9
40.1
49.9
50.6
44.5
45.3
15.3
132.6
129.0
133.1
91.1
88.7
110.8
111.3
101.7
100.6
115.2
115.2
32.7
43.5
43.7
10.3
10.2
11.4
11.3
11.3
11.4
34.8
34.9
35.1
62.2
61.2
61.3
62.5
37.3
37.3
37.3
37.1
37.0
37.1
39.9
40.3
40.2
40.3
40.3
40.2

L crd * approximated in ABAQUS

145

M crd,DH
k*in
50.3
17.5
--115.7
110.0
28.3
28.5
28.1
28.5
--------43.0
47.0
10.0
10.0
99.5
91.7
35.7
35.3
19.7
19.6
19.6
--------58.8
59.7
46.6
47.4
--130.8
120.7
133.2
91.4
88.6
------------------------------53.6
56.5
56.6
---------------------------------

M crd,D
k*in
41.7
14.6
15.0
103.9
101.8
36.9
37.1
36.4
36.9
234.6
238.1
56.4
46.1
39.9
41.6
12.4
11.9
113.5
100.5
39.1
40.4
24.6
23.3
23.0
15.9
15.8
58.3
58.6
73.2
77.5
59.2
66.7
28.1
128.3
120.3
130.2
90.2
87.2
152.3
153.3
137.1
136.8
125.6
126.2
46.7
76.7
75.9
16.8
16.7
30.4
27.0
28.2
28.3
62.6
62.6
63.3
71.6
70.6
70.3
73.6
47.2
47.2
47.2
47.0
45.9
47.1
51.7
53.3
53.2
53.5
53.6
53.6

L crd *
in.
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
14.0
14.0
14.0
14.0
12.0
12.0
12.0
14.0
14.0
16.0
16.0
25.0
25.0
12.0
12.0
20.0
23.0
24.0
27.0
12.0
6.0
6.0
5.0
5.0
6.0
12.0
12.0
12.0
10.0
13.0
12.0
10.0
27.0
24.0
10.0
10.0
17.0
12.0
14.0
13.0
24.0
24.0
24.0
11.0
11.0
11.0
11.0
12.0
12.0
12.0
12.0
12.0
12.0
11.0
12.0
12.0
12.0
12.0
12.0

M cr l

M crd

k*in
67.56
17.41
17.29
192.81
192.98
38.82
38.07
37.81
38.28
279.02
268.64
47.71
45.90
57.15
57.44
9.95
9.96
149.82
150.47
36.63
36.36
20.17
20.08
20.09
10.41
10.24
39.38
39.39
49.97
49.16
44.76
44.85
14.93
129.97
130.42
123.80
86.93
87.44
107.87
108.35
99.22
100.03
118.25
118.89
31.96
41.15
41.02
9.86
9.99
11.06
11.11
11.07
11.05
33.85
34.45
34.43
63.23
61.49
62.74
61.83
36.39
36.39
36.39
36.12
36.44
36.49
38.88
38.90
39.48
39.17
38.72
39.43

k*in
38.67
15.11
13.02
108.11
103.74
39.60
37.12
33.96
40.56
259.43
250.06
51.22
45.66
47.86
48.17
12.96
13.09
114.50
115.07
41.75
41.23
24.92
24.86
24.91
20.76
20.28
58.49
58.87
65.68
78.83
64.75
68.02
28.30
128.02
128.44
116.30
85.59
95.68
151.94
152.82
140.34
141.76
137.30
143.78
47.69
78.20
77.69
17.06
17.28
34.95
33.01
32.28
33.66
64.10
64.07
65.10
72.18
71.40
68.34
73.80
47.48
47.48
47.48
47.00
47.68
47.73
52.75
52.84
55.34
53.34
52.57
55.28

CUFSM elastic buckling, no hole

Channel 2
L crd *
in.
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
15.0
15.0
15.0
15.0
20.0
20.0
12.0
12.0
19.0
19.0
24.0
24.0
12.0
6.0
6.0
6.0
5.0
7.0
11.0
11.0
11.0
12.0
24.0
24.0
11.0
24.0
24.0
10.0
11.0
18.0
18.0
16.0
16.0
26.0
26.0
24.0
11.0
11.0
11.0
12.0
11.0
11.0
11.0
12.0
12.0
12.0
11.0
11.0
11.0
12.0
12.0
12.0

M cr l

M crd

k*in
67.97
17.26
17.27
192.81
193.09
38.46
38.07
37.48
38.22
271.46
260.27
49.93
48.11
57.01
58.26
9.90
9.92
149.82
150.13
36.63
36.36
20.17
20.08
20.09
10.41
10.34
39.38
39.64
49.25
49.91
44.05
44.85
15.07
127.47
124.20
128.76
88.37
85.87
107.98
108.50
99.22
98.12
112.42
112.38
31.88
42.37
42.50
10.02
9.95
11.08
10.96
11.03
11.08
34.64
33.93
34.16
63.49
62.15
62.41
63.79
36.39
36.39
36.39
36.12
36.09
36.13
38.88
39.28
39.19
39.17
39.18
39.14

k*in
41.93
14.63
15.02
105.69
103.74
37.25
37.83
37.25
37.44
239.70
243.99
57.30
50.85
46.82
53.24
12.96
12.75
114.50
101.52
41.75
40.72
25.40
25.31
24.91
20.76
20.48
58.49
58.87
77.15
85.21
63.81
68.02
28.30
126.88
116.81
128.47
96.81
91.40
151.94
152.82
140.34
141.76
126.65
127.17
47.69
78.20
77.69
17.06
17.28
34.95
31.93
32.71
33.66
64.10
64.07
65.10
72.30
70.50
68.34
73.80
47.48
47.48
47.48
47.00
47.03
47.09
52.75
53.55
53.43
53.34
53.44
53.37

Channel 1
L crd *
in.
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
15.0
15.0
15.0
15.0
20.0
20.0
12.0
12.0
19.0
19.0
24.0
24.0
12.0
6.0
6.0
6.0
5.0
7.0
11.0
11.0
11.0
12.0
14.0
14.0
11.0
24.0
24.0
10.0
11.0
18.0
18.0
16.0
16.0
26.0
26.0
24.0
11.0
11.0
11.0
12.0
11.0
11.0
11.0
12.0
12.0
12.0
11.0
11.0
11.0
12.0
12.0
12.0

Channel 2

M cr l

M crd

L crd

M cr l

M crd

L crd

k*in
67.34
17.23
17.13
192.63
192.17
38.47
38.02
37.81
38.26
280.76
276.49
48.73
47.84
57.50
57.27
9.85
9.85
152.67
151.76
36.82
36.62
19.95
19.84
19.92
10.43
10.43
39.13
39.16
49.74
48.80
44.52
44.52
14.82
127.62
128.58
126.98
85.18
87.17
107.41
107.91
98.79
99.81
118.12
118.74
31.78
40.75
40.58
9.80
9.94
11.00
11.07
11.01
10.99
34.47
34.29
34.19
62.52
60.45
61.98
60.88
36.22
36.22
36.22
35.95
35.95
35.95
38.71
39.10
39.03
38.99
38.92
38.99

k*in
36.38
14.46
12.57
105.73
100.37
34.97
33.32
31.56
35.31
236.79
247.64
50.74
46.59
37.15
37.08
9.82
9.82
85.95
85.53
30.06
29.50
17.75
17.67
17.71
14.39
14.18
42.30
42.28
54.18
52.00
47.53
47.37
19.61
122.79
124.46
108.98
83.61
87.94
123.59
124.67
116.37
118.65
108.07
111.45
41.89
50.55
49.76
15.47
16.75
23.66
23.97
23.66
23.64
45.07
45.48
44.07
71.58
67.40
70.20
68.02
46.03
46.03
46.03
45.72
45.72
45.72
51.31
51.56
51.52
51.50
51.46
51.50

in.
8.90
12.90
10.80
10.80
10.90
16.10
15.70
15.70
15.80
11.50
11.50
11.40
13.90
10.60
10.60
15.10
15.10
10.80
10.70
15.50
15.50
15.50
15.50
15.50
22.30
18.40
14.80
14.80
21.20
21.20
25.80
25.80
17.40
9.30
9.30
9.30
9.20
9.30
16.30
16.30
16.20
16.20
19.40
23.40
13.30
23.60
23.60
13.30
13.30
23.50
23.50
23.40
23.40
28.20
28.20
28.30
16.60
16.60
16.60
16.60
13.40
13.40
13.40
13.40
13.40
13.40
13.40
16.20
16.20
16.20
16.20
16.20

k*in
68.45
17.07
17.13
193.11
193.72
37.99
37.70
37.13
37.67
271.13
272.01
49.73
47.84
56.85
58.09
9.77
9.79
149.19
148.57
36.32
35.99
20.05
19.99
19.91
10.26
10.08
39.16
39.37
48.85
49.63
43.55
44.49
14.96
127.00
127.39
128.71
88.03
85.54
107.55
108.08
98.95
97.56
111.09
111.08
31.73
42.16
42.33
9.98
9.91
11.02
10.88
10.96
11.02
33.49
33.56
33.87
62.75
61.49
61.63
63.16
36.22
36.22
36.22
35.95
35.95
35.95
38.71
39.10
39.03
38.99
38.92
38.99

k*in
40.36
14.02
14.38
102.82
100.85
33.35
33.41
32.52
33.19
236.79
240.79
60.03
52.46
36.11
40.76
9.75
9.58
85.39
75.98
29.93
28.88
18.16
18.11
17.72
15.02
14.61
42.28
42.90
50.96
54.97
42.89
47.37
20.78
120.90
109.29
124.56
90.23
84.53
124.50
124.73
115.58
112.93
83.24
83.37
41.80
56.35
56.95
16.63
16.19
23.69
22.79
23.38
23.93
42.24
42.15
44.53
72.24
69.30
69.43
73.02
46.03
46.03
46.03
45.72
45.72
45.72
51.31
51.56
51.52
51.50
51.46
51.50

in.
10.80
12.80
13.10
10.80
10.90
15.70
15.60
15.90
15.80
11.50
11.50
13.80
11.40
10.60
12.70
15.10
15.20
12.90
10.70
15.60
15.60
15.50
15.50
15.50
22.30
22.30
14.80
14.70
17.60
21.30
25.80
25.80
17.40
9.30
9.30
9.30
11.20
9.30
16.30
16.30
16.20
13.50
16.10
16.10
13.30
28.50
28.50
13.30
13.30
23.50
19.40
23.40
23.40
23.40
23.40
23.40
16.60
16.60
16.60
16.60
13.40
13.40
13.40
13.40
13.40
13.40
13.40
16.20
16.20
16.20
16.20
16.20

3.1.6 Influenceofholesonbeamlocalanddistortionalcriticalelasticbucklingloads
135B

4.3.1.6.1 Localbuckling
198B

TheABAQUSlocalbucklingeigenbucklingresultsforeachbeamspecimenCsection

(Channel 1 and Channel 2) with holes is compared to the same beam specimen but
without holes in Figure 4.68. The variation in Mcr for the LH, LH2, and L modes (see
1294H

Section 4.3.1.4.1fordefinition)withholesizetoflatwebdepthishighlightedin Figure


1295H

1296H

4.68a. The LH mode (buckling of the compressed unstiffened strip above a hole) is
observed only when 0.20<hhole/h<0.40, and is always the lowest buckling mode when it
exists. As hhole/h exceeds 0.40 the lowest mode switches to the LH2 mode. This trend
occurs because as h decreases, the local buckling halfwavelength decreases causing
multiple halfwaves to form in the unstiffened strip at the hole. When hhole/h<0.20 the
unstiffenedstripabovetheholeisrelativelystiff(i.e.,deeprelativetoholelength)and
platebucklingcontrolsasthelowestlocalbucklingmode.TheminimumMcrfortheLH,
LH2, and L is plotted in Figure 4.68b exhibits a similar trend to that observed for
1297H

stiffened elements in bending (see Figure 3.26a), where the maximum hole influence
1298H

occurswhenhhole/hisbetween0.30and0.40.Unstiffenedstripbuckling(LHandLH2)of
fullmemberscontrolsforhhole/hexceeding0.50whichisalsoconsistentwiththebehavior
ofastiffenedelementinbending(anddifferentfromacolumnwithholes,whereweb
localbucklingoccursawayfromtheholeforlargehhole/h).ThepresenceoftheCsection
flangesreducesthemagnitudeoftheholeinfluenceinafullmemberwhencomparedto

146

a stiffened element, which is consistent with similar observations for compression


members(seeFigure4.4).
129H

1.5

1.5
LH controls
LH2 controls
L controls

Mcrl,ABAQUS hole/Mcrl,ABAQUS no hole

Mcrl,ABAQUS hole/Mcrl,ABAQUS no hole

LH
LH2
L

0.5

0.1

0.2

0.3

0.4

0.5
hhole/h

0.6

0.7

0.8

0.9

0.5

0.1

0.2

0.3

0.4

0.5
hhole/h

0.6

0.7

0.8

0.9

Figure4.68InfluenceofholesonbeamspecimenMcrl (Channel1andChannel2plotted)considering(a)all
localbucklingmodesand(b)thelowestlocalbucklingmode

4.3.1.6.2 Distortionalbuckling
19B

TheABAQUSdistortionalbucklingeigenbucklingresultsforeachbeamspecimenC

section(Channel1andChannel2)withholesiscomparedtothesamebeamspecimen
but without holes in Figure 4.69. The variation in Mcr for the DH and D modes (see
130H

Section 4.3.1.4.2fordefinition)withholesizetoflatwebdepthishighlightedin Figure


130H

1302H

4.69a. The DH mode is often the lowest distortional mode in Figure 4.69b, especially
130H

whenhhole/hisbetween0.20and0.40.Thismodeisinitiatedbyunstiffenedstripbuckling
and is related to the LH mode, and therefore its maximum influence in this region is
expected.

Theratioofwebdepthtoflangewidthisanimportantparametertoconsiderwhen

differentiatingbetweentheLHandDHmodesforbeamswithholes.TheDHmodeis
most prevalent in the range 2<H/B<6 as shown in Figure 4.70. As the beam depth
1304H

increases relative to flange width (H/B>6) the distortional tendency associated with

147

unstiffened strip buckling decreases and the DH mode transitions to the LH (or LH2)
mode.
1.5

1.5
DH controls
D controls

Mcrd,ABAQUS hole/Mcrd,ABAQUS no hole

Mcrd,ABAQUS hole/Mcrd,ABAQUS no hole

DH
D

0.5

0.1

0.2

0.3

0.4

0.5
hhole/h

0.6

0.7

0.8

0.9

0.5

0.1

0.2

0.3

0.4

0.5
hhole/h

0.6

0.7

0.8

0.9

Figure4.69InfluenceofholesonbeamspecimenMcrd(Channel1andChannel2plotted)asafunctionof
holedepthtoflatwebdepthconsidering(a)alldistortionalbucklingmodesand(b)thelowestdistortional
bucklingmode

1.5

1.5
DH controls
D controls

Mcrd,ABAQUS hole/Mcrd,ABAQUS no hole

Mcrd,ABAQUS hole/Mcrd,ABAQUS no hole

DH
D

0.5

4
H/B

0.5

4
H/B

Figure4.70InfluenceofholesonbeamspecimenMcrd(Channel1andChannel2plotted)asafunctionof
webdepthtoflangewidthconsidering(a)alldistortionalbucklingmodesand(b)thelowestdistortional
bucklingmode

3.1.7 Influence of experiment boundary conditions on beam local and distortional


criticalelasticbucklingloads
136B

4.3.1.7.1 Localbuckling
20B

The influence of experiment boundary conditions on the elastic buckling


behavior is evaluated by comparing the ABAQUS critical elastic buckling moment

148

Mcrl (without holes) of each Csection making up the beam specimens to the local
buckling moment determined with the finite strip software CUFSM. Since the finite
strip method considers elastic buckling of each channel individually under a constant
moment, the comparison of ABAQUS and CUFSM results isolate the influence of the
aluminumanglestrapsatthetopandbottomflanges,aswellasthelateralbracingand
theapplicationoftheconstantmomentasaseriesofpointloadsintheexperiments.The
experiment loading and boundary conditions have a minimal influence on Mcrl for the
specimens considered in this study as shown in Figure 4.71. This result is consistent
1305H

withthelocalbucklingmodeshapesinSection 4.3.1.4.1,whereitwasobservedthatthe
1306H

formationoflocalbucklinghalfwavesintheconstantmomentregionwereunimpeded
bytheanglestraps.

Mcrl,no hole, ABAQUS/Mcrl,CUFSM

1.5

0.5

50

100

150
H/t

200

250

Figure4.71InfluenceoftestboundaryconditionsonMcrl

149

300

4.3.1.7.2 Distortionalbuckling
201B

The influence of the experiment boundary conditions on the distortional


buckling behavior is evaluated by comparing the critical elastic buckling moment Mcrd
(without holes) from the ABAQUS eigenbuckling analyses to the buckling moment of
each channel individually determined with the finite strip software CUFSM. The
comparison of ABAQUS and CUFSM results isolates the influence of the aluminum
angle straps, lateral bracing and the load application method on the critical elastic
moment results. The experiment test conditions provide a significant boost to Mcrd as
shownin Figure4.72b,whichishypothesizedtoberelatedtotherestraineddistortional
1307H

bucklingcausedbythealuminumanglestraps.Thishypothesisissupportedbyexisting
research on unrestrained elastic distortional beam buckling (no compression flange
connections), which observed similar CUFSM and ABAQUS eigenbuckling results (Yu
and Schafer 2006). The pure D distortional halfwavelengths approximated from
ABAQUS(forspecimenswithoutholes)in Figure4.72bareoftenshorterrelativetothe
1308H

predicted halfwavelengths from a finite strip analysis. This trend is a direct result of
the angle spacing (12 on center for Test Sequences 1 and 3, 6 for Test Sequence 2),
whichislessthanthefundamentalLcrdformanyoftheCsections.Thechangeinhalf
wavelength away from the natural halfwavelength of the distortional mode increases
thecriticalelasticbucklingmoment.ThisboostinMcrddecreaseswithincreasingH/Bas
shown in Figure 4.72a, since the fundamental Lcrd also decreases as beam web depth
1309H

increasesrelativetoflangewidth.

150

1.6

1.5

1.2

Lcrd,ABAQUS no hole/Lcrd,CUFSM

Mcrd,no hole, ABAQUS/Mcrd,CUFSM

1.4

1
0.8
0.6
0.4

0.5

0.2
0

4
H/B

4
H/B

Figure4.72Influenceoftestboundaryconditionson(a)Mcrdand(b)onthedistortionalhalfwavelength

The boost in Mcrd from the restraint of the beam compression flanges exhibits a
linear trend when plotted against the ratio of Lcrd (from CUFSM) versus the restraint
spacing Sbrace in Figure 4.73. A linear equation is fit to this trend, resulting in a useful
130H

approximationoftherestraintboost:

L
Dboost = 0.15 crd
S brace

L
+ 0.85 , crd 1
S brace

(4.15)

2
Beam database
Eq. (4.15)
DSM Design Guide

1.8

Mcrd,ABAQUS no hole/Mcrd,CUFSM

1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0

0.5

1.5

2
2.5
3
Lcrd,CUFSM/Sbrace

3.5

4.5

Figure4.73BoostinMcrdfromtheanglerestraintsincreasesasthefundamentaldistortionalhalfwavelength
increasesrelativetotherestraintspacingSbrace

151

TheDSMDesignGuidessuggestedmodificationtoMcrdwhenL<=Lcrdisalsoplottedin
Figure4.73(AISI2006,Section4.2):
13H

*
M crd ( L < Lcrd ,CUFSM ) = M crd
( L Lcrd ,CUFSM )

ln (L Lcrd ,CUFSM

(4.16)

where M*crd is the minimum distortional critical elastic buckling moment read from
CUFSM.LisassumedequaltoSbracewhenplottedinFigure4.73,i.e.thedistortionalhalf
132H

waveisassumedtoformbetweentheflangebraces.TheDSMDesignGuideprediction
forMcrdishigherthanthatproposedbyEq. (4.16)becauseformanyofthebeamsinthe
13H

ABAQUSgenerated elastic buckling database, Lcrd was shortened but not completely
restrainedbetweenbraces.Ontheotherhand,theABAQUSeigenbucklinganalysesdid
not simulate contact between the angles and the flanges (only the bending and shear
stiffness of the angles), and therefore the actual Mcrd most likely lies between the two
predictions.

3.2 Approximate prediction methods for use in design


75B

3.2.1 Localbuckling
137B

4.3.2.1.1 Predictionmethod
20B

The approximate method for predicting the local elastic buckling behavior of cold

formed steel beams is similar to the method for columns presented in Section 4.2.7.1.1
134H

LocalbucklingisassumedtooccurastheminimumofMcr ofthegrosscrosssection(as
calculated in the Direct Strength Method) and local buckling of the compressed
unstiffenedstripadjacenttothehole,Mcrh.Themethodcapturesthelowestunstiffened

152

stripbucklingmode,eithertheLHorLH2mode,withtheproceduredescribedinFigure
135H

4.17. When the hole length is longer than the fundamental halfwavelength of the net
crosssectionLcrh,thentheLH2modegoverns.WhentheholelengthislessthanLcrh,the
LHmodegoverns.

TopredictMcrhfromthenetcrosssectioninCUFSM,thecrosssectionisrestrainedto

isolatelocalbucklingfromdistortionalbucklingasshownin Figure4.74.Compressed
136H

corners should be restrain in the direction normal to the neutral axis about which
bendingoccurs(cornersexperiencingtensionneednotberestrained).Itisimportantto
avoid fully restraining a crosssection element, since this prevents Poissontype
deformations and artificially stiffens the crosssection. The only time both the x and z
directions of a corner should be restrained is if a hole isolates two compressed
intersecting elements (as in the case of a flange hole in a Csection, see Figure 4.74a).
137H

Finally, when holes isolate two compressed elements of a crosssection (similar to the
flangeholeinthecolumnhatsection,see Figure4.16b),theisolatedelementshouldbe
138H

removedfromthecrosssection.Thispredictionmethodisvalidatedinthenextsection
usingthebeamelasticbucklingdatabasedevelopedinSection4.3.1.5.
139H

153

Restrain compressed corners


in direction normal to neutral
axis (typ.)
Compression

z
Compression

Restrain isolated
compressed corners in
x and z

Neutral axis

Tension

Tension

Web hole

Flange hole

a
b

Compression

Compression
Neutral axis

Tension

Tension

Figure4.74GuidelinesforrestrainingbeamnetcrosssectionsintheCUFSMlocalbucklingapproximate
method

4.3.2.1.2 Methodverificationusingelasticbucklingdatabase
203B

The finite strip prediction method is used to predict Mcrl for the 144 Csections

describedin Table4.11.ThesepredictionsarecomparedtotheABAQUSeigenbuckling
1320H

resultsfromTable4.12(theminimumofL,LH,andLH2modes),anddemonstratesthat
132H

thefinitestripapproximatemethodisviableandconservativeoverawiderangeofhole
widthsandbeamdepths.AcleartransitionfromLandLH2bucklingtoLHbuckling
occursastheCsectionsincreasesindepthasshowninFigure4.75a.Thisobservationis
132H

consistent with finite element eigenbuckling observations (see Figure 4.56 to Figure
132H

1324H

4.60), where as beam depth increases the halfwavelength of the netsection increases
beyond the length of the hole, resulting in a switch from unstiffened strip buckling in
twohalfwaves(LH2)toonehalfwave(LH).Themeanandstandarddeviationofthe
ABAQUStopredictedratioforMcrlare1.14and0.16respectively.

154

2
L predicted
LH2 predicted
LH predicted

1.8

1.6

1.4

Mcrl,ABAQUS/Mcrl,predicted

Mcrl,ABAQUS/Mcrl,predicted

1.6

1.2
1
0.8
0.6

1.4
1.2
1
0.8
0.6

0.4

0.4

0.2

0.2

50

100

150
H/t

200

250

L predicted
LH2 predicted
LH predicted

1.8

300

0.1

0.2

0.3

0.4

0.5
hhole/h

0.6

0.7

0.8

0.9

Figure4.75ComparisonofABAQUStopredictedMcrlforCsectionswithholesinthebeamdatabaseasa
functionof(a)webdepthand(b)holewidthrelativetoflatwebdepth

3.2.2 Distortionalbuckling
138B

4.3.2.2.1 Predictionmethod
204B

Theweightedaverageandmechanicsbasedfinitestripmethodsforpredicting

McrdofcolumnswithholesintroducedinSection 4.2.7.2.1areemployedheretopredict
1325H

the distortional critical elastic buckling load of coldformed steel beams with holes.
These approximate methods are evaluated against the ABAQUS Mcrd (the minimum of
theDHandDmodes)fromthebeamexperimentdatabaseinTable4.12.
1326H

4.3.2.2.2 Methodverificationusingelasticbucklingdatabase
205B

Figure 4.76 plots Mcrd determined with ABAQUS versus the predictions using the
1327H

weightedaverage and mechanicsbased approximate methods. The ABAQUS Mcrd


(withholes)ismultipliedbytheratioofMcrdfromCUFSMtoMcrdfromABAQUSwithout
holestoeliminatetheinfluenceoftheboundaryconditionsandtoallowforaconsistent
comparison between the ABAQUS results (with only the hole influence) and the
prediction method. The ABAQUS to mechanicsbased prediction ratio is more

155

accurate (ABAQUS to predicted mean of 1.04and standard deviation of 0.02) than the
weightedaverage prediction (ABAQUS to predicted mean of 1.10 and standard
deviationof0.06),whichisconsistentwiththeverificationstudyforcolumnswithholes
inSection4.2.7.2.1.

Mcrd,ABAQUS/Mcrd,predicted*Mcrd, CUFSM/Mcrd,ABAQUS no hole

1328H

2
weighted average
mechanics-based

1.8
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0

0.1

0.2

0.3

0.4

0.5
hhole/h

0.6

0.7

0.8

0.9

Figure4.76ComparisonofmechanicsbasedandweightedaveragepredictionmethodstoABAQUS
resultsforthedistortionalbucklingloadMcrdofCsectionswithholesintheelasticbucklingdatabase

3.2.3 Globalbuckling
139B

The weighted thickness and weighted properties approximate methods

presented in Section 4.2.7.3.1 are now implemented to predict Mcre for a beam with
1329H

uniformlyspacedholesloadedwithaconstantmoment.

156

4.3.2.3.1 DescriptionofPredictionMethod
206B

TheweightedthicknessandweightedpropertiesapproximatesforIy,J,andCw

are employed with the classical lateraltorsional stability equation to predict Mcre of a
beamwithholes(Chajes1974):

2
M cre = EI y GJ + ECw 2 .
L

(4.17)

4.3.2.3.2 ExampleandVerification
207B

The long SSMA 1200S16268 member evaluated in Section 4.2.7.3.2 as a column is


130H

now evaluated as a beam with a uniform moment along the member to compare
prediction methods to ABAQUS results. The ABAQUS boundary conditions and
appliedloadingaredescribedin Figure4.77. Thebeamendsaremodeledaswarping
13H

freeandthecrosssectionatthelongitudinalmidlineiswarpingfixed.Warpingatthe
memberendsisvisibleinFigure4.77forthislateraltorsionalbucklingmode.
132H

Cross-section dof fixed in 2


and 3 (warping free)

Cross-section dof fixed in 1


and 1 at longitudinal
midline (warping fixed)
2

Warping free end detail

1 (x)

Cross-section dof fixed in 2


and 3 (warping free)

Moment applied as
consistent nodal loads on
cross section (typ.)

Figure4.77ABAQUSboundaryconditionsandappliedloadingforan
SSMA1200S16268beamwithholes(hhole/h=0.50shown)

157

Figure 4.78 demonstrates that both the weighted stiffness and weighted
13H

properties models are accurate predictors of Mcre for hhole/h0.50 in this particular case.
Forhhole/h>0.50,thereductioninpredictionaccuracyoccursbecausetheweightedaverage
approximationsforCwarenotconsistentwiththeactualphysicalbehavior(Jwasshown
tobeconsistentwiththeweightpropertiesmethodforcalculatingsectionproperties
in Section ). If a designer does not know Cw,avg, then using the net section properties
(calculatedwithCUFSM,see Figure4.37)ortheweightedpropertiespredictionwith
134H

Cw,net=0arebothviableoptionsforconservativelypredictingMcre.

Mcre,ABAQUS/Mcre,prediction

1.5

weighted thickness
weighted properties
net section
weighted properties, ABAQUS Cw ,avg
weighted properties, Cw ,net=0

0.5

0.1

0.2

0.3

0.4

0.5
hhole/h

0.6

0.7

0.8

0.9

Figure4.78Comparisonofweightedthicknessandweightedpropertiespredictionmethodsforthe
SSMA1200S16268lateraltorsionalbeambucklingmode.Predictionsusingnetsectionpropertiesarealso
plottedasaconservativebenchmark.

158

Chapter 5
Experiments on cold-formed steel
columns with holes
4B

TheelasticbucklingmodesdiscussedinChapter4andtheirinfluenceontheload
135H

deformation response of coldformed steel columns can be readily observed and


quantified with experiments. In this study, 24 coldformed steel lipped Csection
columnswithandwithoutslottedwebholesaretestedtofailure.Thecolumnlengths
andcrosssectiondimensionsarespecificallychosentoexploretheconnectionbetween
local, distortional, and global elastic buckling modes, ultimate strength, and the
resulting failure mechanisms. The elastic buckling behavior is evaluated for each
specimenwithafiniteelementeigenbucklinganalysis,takingcaretoaccuratelysimulate
the tested boundary conditions and measured specimen dimensions. These elastic
bucklingresultsareusedtoprovideameansofunderstandingthevarieddeformation
responseunderload.Thecolumnsaretestedwithfrictionbearingboundaryconditions
wheretheendsofeachspecimenaremilledflatandparallel,andbeardirectlyagainst

159

steelplatens.Recommendationsaremadetoadviseotherresearchersontheviabilityof
the frictionbearing boundary conditions when testing short and intermediate length
columns.

5.1 Acknowledgements
34B

Thecoldformedsteelcolumntestsdescribedinthischapterwerecompletedwitha

teameffortfromtheindividualsbelow:

EricHarden
LatrobeHallMachineShop
WalterKrug
MarylandHallMachineShop
MichaelFranckowiak
MarylandHallMachineShop

Dr.RachelSangree
JohnsHopkinsPostdoctoralResearcher
JackSpangler
SeniorMechanicalEngineerStructuresLab
NickolayLogvinosky
StructuresLabTechnician
MarioFasano
JohnsHopkinsSenior

RebeccaPierce
JohnsHopkinsFreshman
DawneshiaSanders
BaltimorePolytechnicInstituteSenior
AlexanderPei
HighSchoolIntern

Clark Western Building Systems in Dundalk, MD graciously donated the structural


studs.

5.2 TestingProgram
35B

Twentyfour coldformed steel lipped Csection columns with and without pre

punchedslottedwebholesweretestedtofailure.Theprimaryexperimentalparameters
are column crosssection, column length, and the presence or absence of slotted web
holes. The specimen naming convention, as it relates to the testing parameters, is
definedinFigure5.1.
136H

160

No Holes
362-1-24-NH
362-2-24-NH
362-3-24-NH
SSMA 362S162-33
362-1-48-NH
362-2-48-NH
362-3-48-NH
600-1-24-NH
600-2-24-NH
600-3-24-NH
SSMA 600S162-33
600-1-48-NH
600-2-48-NH
600-3-48-NH

Holes
362-1-24-H
362-2-24-H Short Column
362-3-24-H
362-1-48-H
Intermediate
362-2-48-H
Column
362-3-48-H
600-1-24-H
600-2-24-H Short Column
600-3-24-H
600-1-48-H
Intermediate
600-2-48-H
Column
600-3-48-H

362-1-24-NH
Specimen with holes
(H) or without holes
(NH)
Cross- section type

Specimen
number within
common group
(1,2,3)

Nominal specimen
length, 24 in. or 48 in.

Figure5.1Columntestingparametersandnamingconvention

2.1 Rationale for selecting specimen dimensions


76B

2.1.1 Crosssectiontypes
140B

TwoindustrystandardcrosssectionsfromtheSteelStudManufacturersAssociation

(SSMA2001),362S16233and600S16233,wereevaluatedinthisstudy.The362S16233
crosssectionhasanominalwebwidthof3.62in.,whilethe600S16233webiswiderat
6.00 in. Both sections have a 1.62 in. flange and nominal sheet thickness of 0.0346 in.
SpecificmeasureddimensionsareprovidedinSection5.2.4.
137H

Thebucklinghalfwavelengthsthatformalongthelengthofthespecimensarecross

section dependent, and can be calculated with the semianalytical finite strip method
(FSM)(Schaferanddny2006).FSMassumessimplysupportedboundaryconditions,
and therefore the local and distortional halfwavelengths for the crosssections studied
here,asprovidedinTable5.1,areonlyaguideastotheexpectedhalfwavelengthinthe
138H

fixedfixedtests.TheFSMhalfwavelengthsarestillausefulreferencewhendecidingon
specimenlengths(seeSection5.2.1.2)andidentifyingbucklingmodes(seeSection5.3.2),
139H

1340H

especially as specimen length increases and local and distortional buckling half

161

wavelengths converge to the fundamental (simply supported) halfwavelengths


reportedinTable5.1.
134H

Table5.1FSMlocalanddistortionalbucklinghalfwavelengthsfornominal362S16233and600S16233
crosssections

Cross-section
362
600

Elastic buckling half-wavelength


Local (L)
Distortional (D)
in.
in.
2.8
15.4
4.7
12.2

2.1.2 Columnlengths
14B

Morethan80%ofthetestedspecimenswithholesavailableintheliteraturearestub

columns,asdepictedinthespecimenlengthhistogramoftestedspecimensprovidedin
Figure5.2.(Thehistogramisconstructedwiththespecimensfromtheelasticbuckling
1342H

database inSection 4.2.6.) Stub columns accommodate local buckling halfwaves, but
134H

due to their short length, distortional buckling is typically restrained from forming at
relevantstresslevels.Thespecimenlengthsselectedinthisstudy,a24in.shortcolumn
anda48in.intermediatelengthcolumn,ensurethatatleastonedistortionalhalfwave
and multiple local halfwaves can form along the length of the column (see Table 5.1).
134H

Further,atleastforNorthAmericanpractice,theselectedlengthsaremoretypicalofthe
unbraced length of actual coldformed steel columns in an allsteel design with
bridginginplacetobracethestuds.

162

20

20

Stub column tests

18

16

16

14

14

number of specimens

number of specimens

18

12

Specimens with
holes in this study

10
8
6

12
10
8
6

10

20

30
40
50
column specimen length (in.)

60

70

10
L/H

12

14

16

18

20

Figure5.2Testedlengthsofcoldformedsteelcolumnswithholesasafunctionof(a)columnlengthLand
and(b)LversusouttooutcolumnwidthH

2.1.3 Holetypeandlocation

Oneslottedwebholeislocatedatthemidheightoftheshortcolumntoevaluateits

influence at the midlength of one distortional buckling halfwave. Two slotted web
holesareorientedintheintermediatelengthcolumnswithanindustrystandardspacing
of 24 in. (SSMA 2001). The holes also coincide with the locations where distortional
bucklinghalfwavesareexpectedtohavetheirmaximumdisplacementunderload.A
typical short column and intermediate length column specimen with slotted holes is
providedinFigure5.3.
1345H

163

24 in.

Short column

Intermediate length column

Figure5.3Typicalcolumnspecimenswithslottedholes

2.2 Column test setup


7B

The column tests were performed with the 100 kip capacity twopost MTS
machineshowninFigure5.4.Theuppercrossheadandloweractuatorarefittedwith12
1346H

in. x 12 in. x 1 in. thick chromemoly 4140 steel platens ground flat and parallel. The
column specimens bear directly on the steel platens as they are compressed. Friction
betweenthecolumnendsandthesteelplatensaretheonlylateralforcesthatrestrainthe
column crosssection under load. An MTS load cell (model number 66123A02)
measuredtheappliedcompressiveforceoneachspecimen,andaninternalMTSlength
voltagedisplacementtransducer(LVDT)reportedactuatordisplacement.

164

All column specimens were loaded in displacement control at a constant rate of


0.004inchesperminute.Thisratewasselectedtoensurethatthe3ksiaxialstressper
minuteupperlimitintheSpecificationforstubcolumntestingwouldnotbeexceeded
(AISITS2022001).AnMTS407controllerwasusedtooperatethehydraulicactuator
duringthecompressiontests.

Fixed
Crosshead

Load Cell
Friction-bearing
boundary conditions
(specimen bears
directly on steel
platen)

Position
transducers
(with
magnet tips)

Figure5.4Columntestsetupandinstrumentation

Two Novotechnik T Series position transducers fitted with balljointed magnet


tips measured the eastwest displacements of the specimen flangelip intersections at
columnmidheight.Eachtransducerhasastrokeofsixinchesandispoweredbyone9
volt battery. The battery strengths were checked periodically to ensure that a drop in
batterychargedidnotinfluencethetransducerreadings.Theloadcellandtransducer
readings are transmitted as voltage to a PC fitted with a National Instruments data
acquisitioncard.Thevoltagesarethenconvertedtoforcesanddisplacementswiththe
conversionfactorssummarizedin Table5.2.Alldisplacementconversionfactorswere
1347H

165

determinedbytheauthorwithavoltmeteranddigitalcalipers.The dataisplottedto
the PC screen and recorded in a text file with a custom LabVIEW program (Labview
2005).
Magnetic tip

Figure5.5Novotechnikpositiontransducerwithballjointedmagnetictip

Table5.2Voltageconversionfactorsforcolumntestinstrumentation

Measurement
Tensile Force
Actuator Displacement
West Flange Displacement
East Flange Displacement

Source
MTS Load Cell
MTS Internal LVDT
Novotechnik Position Transducer
Novotechnik Position Transducer

Conversion
1 Volt = 1000 lbf
1 Volt = 0.300 in.
1 Volt = 0.678 in.
1 Volt = 0.678 in.

2.3 Column specimen preparation


78B

All column specimens were cut from 8 ft. structural studs using the Central
Machinery 4 inch metal cutting band saw shown in Figure 5.15. For short columns
1348H

without holes, the whole series of specimens (for example 362124NH, 362224NH,
and 362224NH) was cut from a single 8 ft. structural stud. For all other specimen
types, each specimen was cut from its own individual stud. Tensile coupons for
materialstestingwereobtainedfromtheleftoverstudlengths.

166


Figure5.6CentralMachinerymetalbandsawusedtoroughcutcolumnspecimens

The specimen ends were milled to ensure flat and parallel bearing surfaces for
testing.Theflatnesstoleranceacrossthespecimenendisrecommendedas0.001inches
for stub columns and was adopted as the goal for this study (Galambos 1998a). The
short columns were sidemilled with a Fadel computer numericallycontrolled (CNC)
verticalmillingmachine.TheintermediatelengthcolumnsweretoolongfortheCNC
machine, and were instead sidemilled with a Bridgeport manual milling machine.
Duringinitialtrialsthemillingprocesscausedtroublesomevibrationsofthespecimen.
Thelargeclampingforcesrequired todampenthevibrationalsotendedtomodifythe
shapeoftheCsectionduringthemillingprocess.Unsatisfactoryflatnessresultswere
obtainedinthesetrials,withflatnessvariationsofupto0.010inches.
Themillingprocedurewasimprovedbyencasingthespecimenendsinbismuth
diaphragms before milling as demonstrated in Figure 5.7. The diaphragms preserved
1349H

the undeformed shape of the specimens, dampened vibration during the milling

167

process, and reduced the clamping force required to hold the specimens in place.
Bismuth is a chemical element that is relatively soft compared to steel at room
temperatureandmeltsat158degreesFahrenheit.

Figure5.7362S16233shortcolumnspecimenwithbismuthenddiaphragms

Liquidbismuthwaspouredintocustomwoodformsatthespecimenends.Once
thebismuthwasset,thespecimen(withbismuthenddiaphragms)waspositionedinthe
milling machine (Figure 5.8 through Figure 5.11). Several passes were made until the
1350H

135H

steelcrosssectionandbismuthdiaphragmwereflush.Bothcolumnendsweremilled
without removing the specimen from the milling table to reduce the chances of
unparallel bearing ends. The bismuth diaphragms were removed from the specimen
with a few taps of a wooden mallet and then melted down for use with the next
specimen. The flatness tolerance of 0.001 inches was achieved for all but four
specimens(seeSection5.2.4.4,themaximumoutofflatnesswas+0.003in.).
1352H

168

Figure5.8600S16233shortcolumnspecimenorientedinCNCmachine

Figure5.9Anendmillisusedtopreparethecolumnspecimens

169

Figure5.10Theintermediatelengthspecimenswereendmilledinamanualmillingmachine

Figure5.11Thespecimensareclampedatthewebsonlytoavoiddistortionofthecrosssection

170

2.4 Column specimen measurements and dimensions


79B

2.4.1 Specimenreferencesystemanddimensionnotation
142B

All column dimensions are measured with reference to the orientation of the

specimen in the testing machine. The assumed reference system and specimen
dimensionnotationareprovidedinFigure5.15.
135H

2.4.2 Crosssectionmeasurements
143B

Theouttooutdimensionsoftheweb,flanges,andlipstiffenersweremeasured
withdigitalcalipersandaluminumreferenceplatesatthemidlengthofthespecimens.
The measurement procedure for a typical crosssection is summarized in Figure 5.12
1354H

(specimen setup) and Figure 5.13 (cross section dimensions). The outside corner radii
135H

weremeasuredusingasetofradiusgaugeswith1/32in.increments.Thecrosssection
dimensions,basedontheaverageofthreeindependentmeasurements,areprovidedfor
eachspecimeninTable5.3.
1356H

171

Check levelness of measuring platform with


the angle indicator. The slope
perpendicular to the length of the specimen
should be as close to zero as possible.

Find and mark the longitudinal midline of


the specimen.

Clamp the specimen to the measuring


platform.

Figure5.12Setupprocedureformeasuringspecimencrosssectiondimensions

172

Clamp a beveled aluminum plate to the


flange. Use the veneer caliper to measure
the distance between the edge of the lip and
the outside face of the beveled plate. The
true dimension (D1 or D2) is then found by
subtracting the thickness of the beveled
plate from the veneer caliper reading.

Clamp beveled alumninum plates to the lip


and web, ofsetting them longituinally by
about 1/2 inch. Make sure that the beveled
faces are oriented so that they are touching
the channel.

Use the extension on the veneer caliper to


measure the distance between the outside face
of the lip plate and the inside face of the web
plate. Make sure that the extension is flush with
the flange surface. The true dimension (B1 or
B2) is found by subtracting the the thickness of
the beveled plate from the veneer caliper
reading.

Clamp beveled alumninum plates to each flange, ofsetting them


longituinally by about 1/2 inch. Make sure that the beveled faces are
oriented so that they are touching the channel.
Use the extension on the veneer caliper to measure the distance
between the outside face of one flange plate and the inside face of the
other flange plate. Make sure that the extension is flush with the web
surface. The true dimension H is found by subtracting the the
thickness of the beveled plate from the veneer caliper reading.

Figure5.13Procedureformeasuringspecimencrosssectiondimensions

173

Clamp beveled aluminum plate to flange.

Measure the flange angle with the angle


indicator (F1 and F2).

Clamp the beveled aluminum plate to the


stiffener lip. Measure the flange angle
using the angle indicator (S1 and S2).

Figure5.14Procedureformeasuringflangelipandflangewebangles

The four corner angles of each Csection are measured with a digital angle
indicator as demonstrated in Figure 5.14. The angle indicator has a precision of 0.1
1357H

degrees.TheflangelipanglesS1andS2aremeasuredatthemidlengthofthespecimens;
thewebflangeanglesF1andF2aremeasuredatmultiplepointsalongthespecimenas
denotedin Table5.4.TheCsectioncorneranglemagnitudes,basedontheaverageof
1358H

twoindependentmeasurements,areprovidedforeachspecimeninTable5.4.
1359H

174

Table5.3Summaryofmeasuredcrosssectiondimensions
H
in.
3.654
3.712
3.623
3.583
3.645
3.672
3.624
3.624
3.614
3.622
3.623
3.633
6.037
6.070
6.030
6.040
6.011
6.032
6.018
6.017
6.026
6.010
6.017
6.062

Specimen
362-1-24-NH
362-2-24-NH
362-3-24-NH
362-1-24-H
362-2-24-H
362-3-24-H
362-1-48-NH
362-2-48-NH
362-3-48-NH
362-1-48-H
362-2-48-H
362-3-48-H
600-1-24-NH
600-2-24-NH
600-3-24-NH
600-1-24-H
600-2-24-H
600-3-24-H
600-1-48-NH
600-2-48-NH
600-3-48-NH
600-1-48-H
600-2-48-H
600-3-48-H

North

West

B2

D1

D2

RT1

RT2

RB1

RB2

in.
1.550
1.586
1.677
1.650
1.627
1.674
1.611
1.609
1.604
1.602
1.594
1.604
1.599
1.582
1.601
1.594
1.608
1.606
1.621
1.596
1.585
1.598
1.589
1.632

in.
1.621
1.585
1.679
1.595
1.593
1.698
1.605
1.585
1.599
1.595
1.610
1.610
1.631
1.614
1.591
1.606
1.602
1.577
1.609
1.601
1.627
1.625
1.607
1.588

in.
0.411
0.416
0.425
0.430
0.440
0.418
0.413
0.407
0.425
0.420
0.425
0.395
0.488
0.472
0.369
0.484
0.369
0.360
0.486
0.482
0.489
0.480
0.476
0.366

in.
0.431
0.422
0.399
0.437
0.391
0.426
0.426
0.421
0.401
0.412
0.403
0.432
0.365
0.380
0.483
0.359
0.500
0.478
0.374
0.357
0.338
0.388
0.356
0.480

in.
0.188
0.172
0.188
0.188
0.188
0.188
0.172
0.188
0.188
0.172
0.172
0.172
0.172
0.203
0.156
0.172
0.172
0.172
0.172
0.172
0.172
0.188
0.172
0.172

in.
0.188
0.203
0.172
0.203
0.188
0.188
0.172
0.172
0.188
0.172
0.172
0.172
0.156
0.203
0.172
0.172
0.172
0.172
0.172
0.172
0.172
0.156
0.172
0.172

in.
0.172
0.266
0.281
0.281
0.281
0.266
0.281
0.297
0.266
0.281
0.281
0.281
0.250
0.266
0.266
0.250
0.203
0.250
0.234
0.234
0.266
0.250
0.234
0.219

in.
0.188
0.281
0.281
0.281
0.281
0.266
0.281
0.281
0.266
0.281
0.281
0.250
0.203
0.266
0.219
0.219
0.234
0.203
0.219
0.234
0.219
0.219
0.234
0.250

S1

RT1

B1

S2

D1

RT2

D2

B1

East

B2

F1

F2

hhole

rhole=hhole/2

L
RB1

RB2

Section a-a

Lhole

Hole detail
South

Front View
W1

Section b-b

W2

Figure5.15Specimenmeasurementnomenclature

175

Table5.4Summaryofmeasuredlipflangeandflangewebcrosssectionangles
X
X
S1
S2
F1
F2
X
F1
F2
X
F1
F2
X
F1
F2
X
F1
F2
Specimen
in. degrees degrees in. degrees degrees in. degrees degrees in. degrees degrees in. degrees degrees in. degrees degrees
362-1-24-NH 12 12.767
8.367
6 82.600 84.500 12 86.033 86.833 18 84.533 87.000
362-2-24-NH 12 11.367 11.567 6 86.800 84.800 12 87.600 85.467 18 86.400 83.700
362-3-24-NH 12 9.567
9.433
6 85.700 85.000 12 86.300 85.400 18 85.600 83.000
362-1-24-H 12 11.130 10.930 6 83.200 83.970 12 87.600 85.600 18 84.330 86.430
362-2-24-H 12 4.367
10.267 6 86.000 85.133 12 86.333 85.167 18 84.400 84.500
362-3-24-H 12 10.533 10.833 6 85.200 86.333 12 87.700 86.133 18 87.667 89.033
362-1-48-NH 12 7.800
10.100 12 85.100 85.600 18 84.300 85.000 24 85.000 85.600 30 84.000 85.200 36 85.300 85.700
362-2-48-NH 12 8.000
10.800 12 85.500 84.900 18 84.800 85.100 24 84.200 84.600 30 84.800 85.300 36 85.200 84.900
362-3-48-NH 12 9.100
12.200 12 86.900 84.000 18 85.800 83.900 24 85.300 84.100 30 86.400 83.400 36 86.100 83.700
362-1-48-H 12 8.500
9.800 12 86.500 84.800 18 86.600 85.000 24 85.600 84.200 30 85.500 85.100 36 86.400 84.400
362-2-48-H 12 8.300
11.200 12 86.800 84.800 18 86.500 84.200 24 85.600 83.800 30 85.500 84.100 36 86.700 83.800
362-3-48-H 12 9.700
7.300 12 85.300 85.200 18 84.700 86.100 24 84.100 85.300 30 84.400 84.700 36 85.200 85.000
600-1-24-NH 24 1.567
2.133
6 90.567 92.033 12 92.467 93.733 18 91.433 93.767
600-2-24-NH 24 1.733
2.333
6 91.000 92.033 12 91.167 94.067 18 91.467 93.333
600-3-24-NH 24 -2.167
3.500
6 93.700 89.767 12 94.067 91.033 18 92.733 89.667
600-1-24-H 24 0.967
2.033
6 89.000 91.000 12 90.400 92.267 18 91.200 92.600
600-2-24-H 24 1.800
1.100
6 94.433 90.900 12 93.233 88.733 18 91.967 89.000
600-3-24-H 24 0.100
4.100
6 93.500 90.000 12 93.300 89.300 18 90.100 86.300
600-1-48-NH 24 0.167
1.400 12 91.033 92.933 18 90.833 92.700 24 90.600 92.800 30 91.333 92.900 36 91.667 93.200
600-2-48-NH 24 2.000
2.367 12 90.767 91.900 18 90.233 92.300 24 89.900 91.867 30 90.967 92.000 36 91.467 92.767
600-3-48-NH 24 2.600
2.300 12 90.000 92.100 18 89.200 91.900 24 90.000 92.100 30 90.700 92.600 36 90.900 92.500
600-1-48-H 24 2.533
2.100 12 90.933 92.167 18 91.000 92.767 24 90.000 92.633 30 91.000 92.000 36 91.100 92.967
600-2-48-H 24 2.400
1.000 12 89.000 90.700 18 89.200 91.000 24 88.900 91.200 30 89.600 91.600 36 90.200 92.200
600-3-48-H 24 0.667
3.633 12 93.067 89.400 18 93.000 89.500 24 92.300 89.433 30 93.467 89.900 36 93.467 89.600
NOTE: X is the longitudinal distance from the south end of the specimen

176

2.4.3 Specimenthickness

14B

Allstructuralstudsweredeliveredbythemanufacturerwithazincoutercoating
applied for galvanic corrosion protection. The total zinc thickness (i.e., summation of
the zinc coating thicknesses applied to each side of the steel sheet) and the base metal
thickness(sheetthicknesswithtotalzinccoatingremoved)definedin Figure5.16were
1360H

measured for each specimen. The total zinc thickness was used to calculate the
centerline crosssection dimensions from the outtoout measurements (see Section
5.2.4.2), which were then input along with the base metal thickness into the nonlinear
136H

finiteelementmodelsdiscussedinChapter7.Thebasemetalthicknesswasalsousedto
1362H

calculatethesteelyieldstressprovidedinSection5.2.5.
136H

tzinc=t1+t2
t1

zinc (typ.)

base metal

t2

tbare

Figure5.16Basemetalandzincthicknessdefinitions

Totalzincthicknessandbasemetalthicknessweremeasuredforeachspecimen
from tensile coupons cut from the west flange, east flange, and web of an untested
section of structural stud. The thickness measurements were made to a precision of
0.0001 inches with a digital micrometer fitted with a thimble friction clutch. The
thickness was determined by averaging five measurements taken within the gauge

177

length of the tensile coupon (see Figure 5.27 for the definition of gauge length). The
1364H

base sheet metal thicknesses tbare,w (web), tbare,f1 (west flange), tbare,f2 (east flange) and
correspondingtotalzinccoatingthicknessestzinc,tzinc,f1,andtzinc,f2aresummarizedforeach
specimeninTable5.5.
1365H

Table5.5Specimenbaresteelandzinccoatingthicknesses
Web
Specimen

tbare,w
in.

tzinc,w
in.

362-1-24-NH
N/M
362-2-24-NH 0.0368
362-3-24-NH
362-1-24-H
0.0390
0.0030
362-2-24-H
0.0368
0.0057
362-3-24-H
0.0394
0.0027
362-1-48-NH 0.0392
0.0025
362-2-48-NH 0.0393
0.0025
362-3-48-NH 0.0389
0.0013
362-1-48-H
0.0391
0.0019
362-2-48-H
0.0390
N/M
362-3-48-H
0.0401
0.0000
600-1-24-NH
N/M
600-2-24-NH 0.0438
600-3-24-NH
600-1-24-H
0.0414
0.0042
600-2-24-H
0.0427
0.0039
600-3-24-H
0.0429
0.0031
600-1-48-NH 0.0434
0.0026
600-2-48-NH 0.0435
0.0017
600-3-48-NH 0.0436
0.0015
600-1-48-H
0.0429
0.0022
600-2-48-H
0.0429
N/M
600-3-48-H
0.0430
N/M
NOTE: N/M Not measured

West Flange
tbare,f1
tzinc,f1
in.
in.

East Flange
tbare,f2
tzinc,f2
in.
in.

0.0415

N/M

0.0372

N/M

0.0391
0.0390
0.0394
0.0393
0.0394
0.0391
0.0393
0.0391
0.0400

0.0034
0.0023
0.0018
0.0020
0.0022
0.0009
0.0017
N/M
0.0000

0.0391
0.0391
0.0394
0.0392
0.0393
0.0390
0.0394
0.0391
0.0397

0.0028
0.0034
0.0026
0.0020
0.0026
0.0017
0.0017
N/M
0.0010

0.0432

N/M

0.0438

N/M

0.0422
0.0384
0.0431
0.0436
0.0430
0.0432
0.0426
0.0428
0.0434

0.0044
0.0084
0.0026
0.0024
0.0024
0.0021
0.0023
N/M
N/M

0.0428
0.0424
0.0430
0.0434
0.0430
0.0433
0.0429
0.0431
0.0430

0.0030
0.0042
0.0036
0.0028
0.0023
0.0020
0.0021
N/M
N/M

The zinc coating was removed by immersing the tensile coupons in a ferric
chloride bath for 100 minutes. The immersion time was determined with a study of
coupon thickness variation over time for the 362224H web and the 600224H west
flangetensilecoupons.Thecouponswereremovedfromtheferricchloridebathevery
10 minutes, cleaned, and then measured. Figure 5.17 demonstrates that the coupon
136H

thicknessconvergestoaconstantvalue,thebasemetalthickness,atapproximately100
minutes.

178

The average zinc coating thickness (i.e., average of tzinc, tzinc,f1, and tzinc,f2) for all
specimens was 0.0026 inches using the ferric chloride method described above.
Specimen coating thickness measurements were also made with a Positest DFT digital
thicknessgauge(www.defelsko.com)whichproducedanaveragecoatingthicknessfor
allspecimensof0.0016in.Atthemicroscopiclevel,thebondingofthezinctothesteel
substrateresultsinagradientfrompurezinctoamixtureofsteelandzinc(Porter1991).
This gradient complicates the identification of the nonstructural thickness of the
galvanic coating. The base thickness and coating thickness determined with the ferric
chloride method (as reported in Table 5.5) are used throughout this thesis. Accurate
1367H

identificationofthenonstructuralandstructuralcontributionsofthegalvaniccoatingis
warrantedasatopicoffutureresearch,especiallysincetheloaddeformationresponse
andultimatestrengtharesensitivetobasemetalthickness.

1.4

362-2-24-H Web Coupon


600-2-24-H West Flange Coupon

coupon thickness/initial thickness

1.2
0.8
1
0.6
0.8
0.4

0.6
0.4

0.2

0.2
0

After 100 minutes, zinc coating (tzinc=t1+t2) has


been removed with ferric chloride solution

0
0

0.2
20

0.4
40

0.6

60
time (minutes)

80

0.8
100

1
120

Figure5.17Removaloftensilecouponzinccoatingasafunctionoftime

179

2.4.4 Specimenendflatnessandlength
145B

Aftereachspecimenwassawcutandmilledflat,theverticalheightgauge(with
a precision of 0.001 inches) shown in Figure 5.18 was used to measure the specimen
1368H

length and flatness. For each specimen, two independent length measurements were
takenateachroundedcornerlocationdescribedin Figure 5.19.Theheightgaugeand
1369H

specimenareplacedonthesamesteeltabletoensurethatallmeasurementsaremadein
thesamereferenceplane.Thesteeltablewascheckedforflatnesswithadialgaugeand
precisionstandbeforemeasurementsproceeded.LengthsLRT1,LRT2,LRB1,andLRB2as
wellastheaveragelengthLareprovidedforeachspecimenin Table5.6.Thespecimen
1370H

flatness,definedasthedifferencebetweenLRT1,LRT2,LRB1,andLRB2andtheaverage
lengthL,isreportedin Table5.7.Allbutfourspecimensmettheflatnesstoleranceof
137H

0.001 inches, with intermediate length column 362248H having the maximum
deviationof+0.003inchesatLRT2.

North

Figure5.18Aheightgaugeisusedtomeasurespecimenlength

180

LRT1

LRT2

West

East

LRB2

LRB1

Figure5.19LengthsaremeasuredatthefourcornersoftheCsectioncolumn

Table5.6Measuredcolumnspecimenlength
Specimen
362-1-24-NH
362-2-24-NH
362-3-24-NH
362-1-24-H
362-2-24-H
362-3-24-H
362-1-48-NH
362-2-48-NH
362-3-48-NH
362-1-48-H
362-2-48-H
362-3-48-H
600-1-24-NH
600-2-24-NH
600-3-24-NH
600-1-24-H
600-2-24-H
600-3-24-H
600-1-48-NH
600-2-48-NH
600-3-48-NH
600-1-48-H
600-2-48-H
600-3-48-H

LRT1
in.
24.100
24.097
24.097
24.100
24.097
24.099
48.214
48.303
48.192
48.217
48.232
48.196
24.100
24.102
24.100
24.102
24.098
24.101
48.255
48.250
48.295
48.089
48.253
48.061

LRT2
in.
24.100
24.098
24.098
24.099
24.099
24.099
48.214
48.300
48.19
48.216
48.232
48.200
24.101
24.104
24.098
24.100
24.099
24.101
48.255
48.250
48.294
48.088
48.251
48.061

181

LRB1
in.
24.098
24.099
24.098
24.098
24.099
24.099
48.214
48.301
48.191
48.216
48.231
48.195
24.099
24.102
24.099
24.100
24.100
24.101
48.255
48.250
48.295
48.089
48.253
48.059

LRB2
in.
24.099
24.099
24.099
24.100
24.100
24.100
48.214
48.298
48.189
48.216
48.231
48.198
24.099
24.103
24.099
24.101
24.100
24.100
48.255
48.251
48.294
48.088
48.253
48.059

L (avg.)
in.
24.099
24.098
24.098
24.099
24.099
24.099
48.214
48.301
48.191
48.216
48.232
48.197
24.100
24.103
24.099
24.101
24.099
24.101
48.255
48.250
48.295
48.089
48.253
48.060

Table5.7Specimenendflatness
Flatness (Deviation from Average Length)
Specimen
LRT1
LRT2
LRB1
LRB2
in.
in.
in.
in.
362-1-24-NH
0.001
0.001
-0.001
0.000
362-2-24-NH
-0.001
0.000
0.001
0.001
362-3-24-NH
-0.001
0.000
0.000
0.001
362-1-24-H
0.001
0.000
-0.001
0.001
362-2-24-H
-0.002
0.000
0.000
0.001
362-3-24-H
0.000
0.000
0.000
0.001
362-1-48-NH
0.000
0.000
0.000
0.000
362-2-48-NH
0.002
-0.001
0.001
-0.002
362-3-48-NH
0.002
-0.001
0.001
-0.002
362-1-48-H
0.001
0.000
0.000
0.000
362-2-48-H
0.001
0.001
0.000
0.000
362-3-48-H
-0.001
0.003
-0.002
0.001
600-1-24-NH
0.000
0.001
-0.001
-0.001
600-2-24-NH
-0.001
0.001
-0.001
0.000
600-3-24-NH
0.001
-0.001
0.000
0.000
600-1-24-H
0.001
-0.001
-0.001
0.000
600-2-24-H
-0.001
0.000
0.001
0.001
600-3-24-H
0.000
0.000
0.000
-0.001
600-1-48-NH
0.000
0.000
0.000
0.000
600-2-48-NH
0.000
0.000
0.000
0.001
600-3-48-NH
0.001
-0.001
0.001
-0.001
600-1-48-H
0.001
0.000
0.001
0.000
600-2-48-H
0.001
-0.001
0.001
0.001
600-3-48-H
0.001
0.001
-0.001
-0.001

2.4.5 Locationanddimensionsofslottedholes
146B

The length and width of the slotted holes, Lhole and hhole, were measured to a
precision of 0.001 inches with digital calipers. The eastwest locationsof theholes, W1
and W2, were measured by clamping aluminum plates to the outside surface of the
flangesandthenusingthecaliperextensiontomeasurethedistancefromtheedgeofthe
hole to the aluminum plate. (This process is similar to the crosssection measurement
proceduresdescribedinFigure5.13.)Theholesizeandweblocationdimensions,based
1372H

ontheaverageofthreeindependentmeasurements,areprovidedforeachspecimenin
Table5.8.
137H

182

Table5.8Measuredslottedholedimensionsandlocations
Specimen
362-1-24-H
362-2-24-H
362-3-24-H
362-1-48-H
362-2-48-H
362-3-48-H
600-1-24-H
600-2-24-H
600-3-24-H
600-1-48-H
600-2-48-H
600-3-48-H

X
in.
L/2
L/2
L/2
(L-24)/2
(L-24)/2
(L-24)/2
L/2
L/2
L/2
(L-24)/2
(L-24)/2
(L-24)/2

W1
in.
0.946
1.146
0.935
1.252
1.126
0.982
2.147
2.365
2.347
2.161
2.166
2.371

W2
in.
1.141
0.967
1.114
0.974
1.016
1.112
2.361
2.155
2.166
2.375
2.351
2.162

L hole
in.
4.003
4.000
4.005
3.999
4.001
4.000
4.002
4.001
4.001
4.002
4.001
3.999

h hole
in.
1.492
1.502
1.493
1.500
1.496
1.493
1.498
1.491
1.493
1.494
1.499
1.497

X
in.

W1
in.

W2
in.

L hole
in.

h hole
in.

(L+24)/2 1.198 0.952 4.001 1.494


(L+24)/2 1.171 0.973 4.003 1.494
(L+24)/2 0.967 1.133 4.003 1.491

(L+24)/2 2.162 2.383 3.998 1.497


(L+24)/2 2.176 2.360 4.002 1.498
(L+24)/2 2.365 2.156 4.003 1.494

2.4.6 Webimperfections
147B

Variationsinthespecimenwebsweremeasuredtoprovideabasisforthelocal
buckling initial imperfection magnitudes in the specimen nonlinear finite element
modelsconstructedinSection 7.2.Themeasurementsetupshownin Figure5.20usesa
1374H

1375H

dialgaugewithaprecisionof0.001inchesmountedtoalaboratorystandincontactwith
aflatsteeltable.Thespecimenwassupportedhorizontallyatbothendsbyamatching
pairofsteelbarsthatweregroundflatandparallel.Thebarswerealsoincontactwith
the steel table, ensuring that the specimen and the dial gauge were in the same
horizontalreferenceplane.Eachspecimenwebwasmarkedwithagridofmeasurement
pointsshownin Figure5.21.Thestandanddialgaugewereshiftedfromgridpointto
1376H

gridpointandelevationmeasurementswererecorded.Thevariationsfromtheaverage
elevation of the specimen web, based on an average of two measurements per grid
point,areprovidedforeachspecimeninTable5.9.
137H

183


Figure5.20Adialgaugeandprecisionstandareusedtomeasureinitialwebimperfections

+ variation

West

1.2 inches (362 specimens)


2.3 inches (600 specimens)

Center

East

Section a-a

X
CL Web (typ.)

North
a

6 in. (typ.)

Plan view
(short and intermediate length web grid layouts)
Figure5.21Webimperfectionmeasurementgridandcoordinatesystem

184

Table5.9Initialwebimperfections(deviationsfromtheaverageelevationoftheweb)
Specimen
362-1-24-NH

362-2-24-NH

362-3-24-NH

362-1-24-H

362-2-24-H

362-3-24-H

362-1-48-NH

362-2-48-NH

362-3-48-NH

362-1-48-H

362-2-48-H

362-3-48-H

600-1-24-NH

600-2-24-NH

600-3-24-NH

600-1-24-H

600-2-24-H

600-3-24-H

600-1-48-NH

600-2-48-NH

600-3-48-NH

600-1-48-H

600-2-48-H

600-3-48-H

X Distance
West
Center
East
West
Center
East
West
Center
East
West
Center
East
West
Center
East
West
Center
East
West
Center
East
West
Center
East
West
Center
East
West
Center
East
West
Center
East
West
Center
East
West
Center
East
West
Center
East
West
Center
East
West
Center
East
West
Center
East
West
Center
East
West
Center
East
West
Center
East
West
Center
East
West
Center
East
West
Center
East
West
Center
East

in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.
in.

0
0.013
0.022
0.013
0.019
0.015
0.015
0.016
0.017
0.016
0.006
0.016
0.009
0.007
0.014
0.025
0.016
0.021
0.017
0.003
0.009
0.015
-0.008
-0.004
-0.005
0.006
-0.001
0.010
-0.006
-0.015
0.009
0.013
0.011
0.010
0.013
0.019
0.014
0.016
0.055
0.016
0.013
0.061
0.021
0.007
0.034
0.017
0.005
0.052
0.020
0.009
0.020
0.014
0.006
0.007
0.007
0.023
0.060
0.024
0.019
0.060
0.014
0.026
0.055
0.013
0.014
0.059
0.009
0.032
-0.033
0.011
0.017
0.028
0.018

6
-0.007
-0.005
-0.007
-0.006
-0.014
-0.009
-0.004
-0.015
-0.010
-0.008
-0.010
-0.008
-0.009
-0.014
-0.001
-0.009
-0.009
-0.002
-0.010
0.004
0.019
-0.023
-0.016
-0.008
-0.002
-0.003
0.007
-0.003
0.003
0.021
-0.007
-0.011
-0.002
-0.007
-0.005
-0.002
-0.012
0.005
-0.003
-0.019
0.004
0.002
-0.016
-0.023
-0.021
-0.015
0.003
-0.003
-0.014
-0.018
-0.015
-0.001
-0.009
-0.014
-0.003
0.016
0.006
-0.004
0.012
-0.001
-0.003
0.013
-0.003
-0.004
0.012
0.003
0.002
0.023
0.008
0.012
0.003
-0.010

12
-0.011
-0.022
-0.013
-0.010
-0.020
-0.015
-0.010
-0.023
-0.016
-0.014
Hole
-0.013
-0.020
Hole
-0.017
-0.020
Hole
-0.015
-0.014
-0.006
0.010
-0.021
-0.021
-0.011
0.003
0.002
0.006
-0.003
Hole
0.016
-0.003
Hole
-0.004
-0.012
Hole
-0.010
-0.029
-0.027
-0.014
-0.033
-0.030
-0.010
-0.018
-0.029
-0.028
-0.031
Hole
-0.018
-0.024
Hole
-0.027
-0.003
Hole
-0.022
-0.018
-0.010
-0.002
-0.016
-0.005
0.002
-0.021
-0.012
-0.010
-0.026
Hole
0.000
-0.020
Hole
0.004
0.012
Hole
-0.023

Local Variations in Web


18
24
30
-0.004
0.015
-0.013
0.015
-0.004
0.014
-0.006
0.015
-0.007
0.024
-0.008
0.014
-0.003
0.015
-0.003
0.025
-0.008
0.014
-0.001
0.016
-0.009
0.009
-0.001
0.015
-0.003
0.014
-0.007
0.010
-0.009
0.014
-0.010
0.016
-0.015
0.015
-0.002
0.015
-0.011
-0.009
-0.005
-0.006
-0.005
-0.007
0.005
0.004
-0.001
-0.013
-0.002
0.006
-0.015
0.000
0.006
-0.009
0.004
0.010
0.005
0.003
0.002
0.001
-0.002
-0.004
0.003
-0.001
-0.007
0.002
0.007
0.007
0.001
0.006
-0.002
0.010
0.007
-0.004
-0.001
0.000
-0.004
-0.006
-0.003
-0.010
-0.001
-0.003
-0.004
-0.006
-0.003
0.003
-0.010
-0.002
-0.003
-0.006
0.001
0.001
-0.024
0.013
-0.023
0.027
-0.014
0.009
-0.027
0.009
-0.024
0.031
-0.011
0.014
-0.003
0.010
0.006
0.057
-0.012
0.018
-0.019
0.011
-0.017
0.021
-0.012
0.006
-0.011
0.012
-0.001
0.051
-0.013
0.016
-0.006
0.014
-0.002
0.040
-0.016
0.004
-0.026
-0.026
-0.020
-0.018
-0.010
-0.006
-0.008
-0.001
0.003
-0.016
-0.020
-0.025
-0.004
-0.007
-0.009
0.005
0.000
-0.004
-0.021
-0.014
-0.011
-0.011
0.003
0.003
-0.010
0.002
0.008
-0.026
-0.026
-0.024
-0.007
0.002
-0.009
0.002
0.004
0.002
-0.028
-0.022
-0.012
-0.001
0.006
0.005
0.004
0.004
0.007
0.010
0.000
0.007
-0.004
-0.022
-0.014
-0.032
-0.048
-0.040

185

36

42

48

-0.005
-0.005
-0.002
0.015
0.011
0.013
0.000
-0.011
-0.013
0.007
Hole
-0.016
-0.009
Hole
-0.012
-0.003
Hole
-0.007

-0.002
-0.008
-0.005
0.021
0.012
0.016
0.011
-0.001
-0.011
0.010
-0.009
-0.022
-0.002
-0.004
-0.006
-0.004
-0.008
-0.002

0.018
0.013
0.004
0.023
-0.002
0.012
0.007
0.004
-0.001
0.000
-0.009
-0.015
0.016
0.022
0.012
0.017
0.015
0.012

-0.019
-0.009
0.004
-0.024
-0.009
0.003
-0.015
-0.008
-0.002
-0.025
Hole
0.001
-0.019
Hole
0.000
0.009
Hole
-0.045

-0.014
-0.005
0.001
-0.014
-0.006
0.002
-0.015
-0.006
-0.002
-0.007
-0.002
0.004
-0.022
0.002
-0.004
0.011
0.018
0.002

0.016
0.030
0.011
0.011
0.023
0.008
0.013
0.031
0.008
0.009
0.024
0.002
0.013
0.025
0.004
0.020
0.046
0.021

2.4.7 Specimenorientationinthetestingmachine
148B

When placing the specimen in the testing machine, the southern end of the
specimenwasorientedatthebottomplatensuchthatthecenterofthecompressiveforce
was applied through the gross centroid of the Csection. The centerline of the web is
positionedinlinewiththecenterlineofthebottomplatenandoffsettowardsthebackof
thetestingmachineasdescribedin Figure5.22.Thecentroidlocationswerecalculated
1378H

using the centerline dimensions of a nominal SSMA 362S16233 and 600S16233 cross
section.
CL Platen and Column Web

0.380 in. (600S162-33)


0.502 in. (362S162-33)

Location of interior
web edge
CL Platen

Column specimen
Center of platen, center
of load, centroid of Cee
channel

FRONT OF MTS MACHINE

Plan View
(Bottom Platen)

Figure5.22Columnspecimenalignmentschematic

186

The actual cross section and thickness measurements produced centroid offsets
slightly different from the nominal offsets considered in the column tests. The
difference between the nominal and measured offsets, defined here as CG, are
providedin Table5.10.CGproducesendmomentsinthespecimensthatareseveral
1379H

ordersofmagnitudesmallerthantheappliedloadsinthisstudy.Forexample,theend
momentscreatedbyaCGof0.059inchesforspecimen600324NHarecalculatedas
2.0x106kipinchesatpeakload(Ptest=12.24kips)usingthestructuralanalysisprogram
MASTAN(ZiemianandMcGuire2005).TheassumedMASTANstructuralsystemin
Figure 5.23 demonstrates that relatively stiff compression platens and fixedfixed end
1380H

conditionseffectivelyeliminateendmomentsfromsmallloadeccentricities.
Table5.10Specimengrosscentroidandoffsetfromappliedloadduringtests
Specimen

362-1-24-NH
362-2-24-NH
362-3-24-NH
362-1-24-H
362-2-24-H
362-3-24-H
362-1-48-NH
362-2-48-NH
362-3-48-NH
362-1-48-H
362-2-48-H
362-3-48-H
600-1-24-NH
600-2-24-NH
600-3-24-NH
600-1-24-H
600-2-24-H
600-3-24-H
600-1-48-NH
600-2-48-NH
600-3-48-NH
600-1-48-H
600-2-48-H
600-3-48-H

Specimen
Measurements
xcg
tz
in.
in.
0.482
0.038
0.471
0.038
0.504
0.038
0.511
0.042
0.490
0.042
0.524
0.042
0.475
0.041
0.468
0.042
0.475
0.040
0.470
0.041
0.470
0.042
0.486
0.040
0.354
0.047
0.347
0.047
0.344
0.047
0.363
0.046
0.368
0.047
0.361
0.046
0.362
0.046
0.355
0.045
0.353
0.045
0.362
0.045
0.352
0.046
0.356
0.046

Centroid Shift
xcg - tz
used in tests
in.
in.
0.463
0.502
0.452
0.502
0.485
0.502
0.489
0.502
0.469
0.502
0.503
0.502
0.454
0.502
0.447
0.502
0.455
0.502
0.449
0.502
0.449
0.502
0.466
0.502
0.330
0.380
0.323
0.380
0.321
0.380
0.340
0.380
0.344
0.380
0.338
0.380
0.339
0.380
0.333
0.380
0.330
0.380
0.340
0.380
0.329
0.380
0.333
0.380

tz sheet thickness with zinc coating


CS difference measured and as tested centroid offsets

187

CS
in.
0.039
0.050
0.017
0.013
0.033
-0.001
0.048
0.055
0.047
0.053
0.053
0.036
0.050
0.057
0.059
0.040
0.036
0.042
0.041
0.047
0.050
0.040
0.051
0.047

CG
All translation
and rotational
DOF restrained

EI flexural rigidity
1000EI

0.1EI

No moment in
column for stiff
platen even
with load offset

EI

EI
Column specimen
(Centroid shown)
CL Applied
Load

Horizontal
translation and
rotational DOF
restrained
Platen
(typ.)

1000EI

Actuator load

0.1EI

Stiff platen

Structural System

Flexible platen

Moment Diagrams

Figure5.23Influenceofplatenbendingstiffnessonendmomentsforafixedfixedeccentriccolumn

Oncethespecimenisalignedonthebottomplaten,500lbsofcompressiveforce
wasappliedtothecolumnandweakaxisoutofstraightnessmeasurementsweretaken.
The distance from the front of the top and bottom platens to the interior web edge is
denotedasStopandSbottomin Figure5.24.StopandSbottomareobtainedastheaverageofthree
138H

independent measurements with digital calipers as shown in Figure 5.25 and then
1382H

corrected for a systematic platen offset (see Figure 5.24) and the initial web
138H

imperfections in Table 5.9. The initial outofstraightness S provided in Table 5.11 is


1384H

1385H

188

calculatedfromStopandSbottomandimplementedasaninitialgeometricimperfectioninto
thenonlinearfiniteelementmodelsin7.2.
1386H

CL Load
Stop

a
CL Platen

Front of MTS
Machine

Stop
CL Platen

Column Specimen
(orientation exaggerated)
Platen Offset=0.084 in.

Section a-a

S (negative magnitude shown)


Sbottom
Side View
(Looking west)

Figure5.24Columnspecimenweakaxisoutofstraightnessschematic

189


Figure5.25Digitalcalipersareusedtomeasurethedistancefromthecolumnwebtoplatenedge

Table5.11Summaryofoutofstraightnesscalculations
Specimen

Sbottom

Stop

Platen Offset

Sbottom

Correction

Sbottom

Correction

Stop

Corrected for
Corrected for
Top platen edge is Corrected for
Web
Web
As
As
Web
Web
Initial out of
offset from bottom
top platen
Imperfection @
Imperfection
measured measured
Imperfection
Imperfection straightness
platen Edge
offset
X=0
@ X=L
@ X=0
@ X=L

362-1-24-NH
362-2-24-NH
362-3-24-NH
362-1-24-H
362-2-24-H
362-3-24-H
362-1-48-NH
362-2-48-NH
362-3-48-NH
362-1-48-H
362-2-48-H
362-3-48-H
600-1-24-NH
600-2-24-NH
600-3-24-NH
600-1-24-H
600-2-24-H
600-3-24-H
600-1-48-NH
600-2-48-NH
600-3-48-NH
600-1-48-H
600-2-48-H
600-3-48-H

in.
6.507
6.523
6.531
6.524
6.532
6.529
6.352
6.535
6.537
6.530
6.534
6.532
6.352
6.365
6.451
6.356
6.360
6.355
6.346
6.354
6.354
6.311
6.352
6.348

in.
6.622
6.612
6.585
6.613
6.578
6.629
6.393
6.649
6.614
6.554
6.617
6.616
6.472
6.560
6.494
6.486
6.399
6.403
6.436
6.488
6.463
6.458
6.422
6.430

in.
0.084
0.084
0.084
0.084
0.084
0.084
0.084
0.084
0.084
0.084
0.084
0.084
0.084
0.084
0.084
0.084
0.084
0.084
0.084
0.084
0.084
0.084
0.084
0.084

in.
6.591
6.607
6.615
6.608
6.616
6.613
6.436
6.619
6.621
6.614
6.618
6.616
6.436
6.449
6.535
6.440
6.444
6.439
6.430
6.438
6.438
6.395
6.436
6.432

190

in.
0.015
0.015
0.017
0.016
0.014
0.021
0.009
-0.004
-0.001
-0.015
0.011
0.019
0.055
0.061
0.034
0.052
0.020
0.007
0.060
0.060
0.055
0.059
-0.033
0.028

in.
6.577
6.593
6.598
6.592
6.602
6.592
6.427
6.623
6.622
6.629
6.607
6.598
6.381
6.388
6.501
6.388
6.424
6.432
6.370
6.377
6.383
6.336
6.469
6.404

in.
0.022
0.024
0.025
0.009
0.010
0.015
0.013
-0.002
0.004
-0.009
0.022
0.015
0.027
0.031
0.057
0.021
0.051
0.040
0.030
0.023
0.031
0.024
0.025
0.046

in.
6.600
6.588
6.560
6.604
6.568
6.615
6.380
6.651
6.610
6.563
6.594
6.601
6.444
6.529
6.437
6.466
6.348
6.363
6.406
6.465
6.432
6.433
6.396
6.384

in.
-0.024
0.004
0.038
-0.012
0.034
-0.023
0.047
-0.028
0.012
0.066
0.013
-0.003
-0.063
-0.141
0.063
-0.078
0.076
0.069
-0.036
-0.087
-0.049
-0.098
0.072
0.020

2.5 Materials testing


80B

Tensile coupon tests were performed to obtain the steel stressstrain curve and
yieldstressfortheweb,westflange,andeastflangeofeachspecimeninthisstudy.The
tests were conducted in accordance with ASTM specification E 8M04, Standard Test
MethodsforTensionTestingofMetallicMaterials(Metric)(ASTM2004).

2.5.1 Tensilecouponpreparation
149B

Tensilecouponswerealwaysobtainedfromthesame8ft.structuralstudwhich
producedthecolumnspecimen.Flatportionsofthewebandflangeswerefirstrough
cutwithametalbandsawasshownin Figure5.26,andthenfinishedtothedimensions
1387H

in Figure 5.27 with a CNC milling machine. The special jig in Figure 5.27 allowed for
138H

1389H

threetensilecouponstobeproducedatonce.Thetensilecouponswerestrippedoftheir
zinc coating (see Section 5.2.4.3 for procedure) and then measured within the gauge
1390H

length for bare metal thickness, t, and minimum width, wmin. The minimum width is
determinedbytakingtheminimumoffiveindependentmeasurementswithinthegauge
lengthofthespecimenwithdigitalcalipers.

191


Figure5.26Tensilecouponsarefirstroughcutwithametalbansaw

0.38 in.

0.38 in.

1.97 in.

3.18 in.

1.97 in.

0.492 in. *
0.79 in.

R=0.55 in.

gauge length
1.97 in.

*nominal, actual dimension will vary slightly

Figure5.27TensilecoupondimensionsasenteredintheCNCmillingmachinecomputer

192


Figure5.28AcustomjigallowsthreetensilecouponstobemilledatonceintheCNCmachine

2.5.2 Tensiletestsetup
150B

A screwdriven ATS 900 testing machine with a maximum capacity of 10 kips


was used to apply the tensile load. Tensile coupons were positioned in the machine
withfrictiongripsasshownin Figure5.29.Abubblelevelwithashort,straightedge
139H

was used to ensure that each specimen was aligned vertically between the grips. An
MTS 634.11D54 extensometer measured engineering strain and an MTS load cell
measuredforceonthespecimen.Theextensometerwasplacedattheverticalmidlength
of the specimen, centered within the gauge length. The raw voltage data from the
extensometer and load cell were sent to a PC containing a National Instruments data
acquisitioncard.Thevoltagedatawasconvertedtotensileforceandengineeringstrain
usingtheconversionfactorsprovidedin Table5.12.Thedatawasplottedonthescreen
1392H

andrecordedtoafilewithacustomLabVIEWprogram(Labview2005).

193

Figure5.29ATSmachineusedtotesttensilecoupons

Table5.12Voltageconversionfactorsfortensilecoupontesting
Tensile Coupon Testing
Measurement

Source

Tensile Force

MTS Load Cell

Conversion
1 Volt = 1000 lbf
-5

Engineering Strain MTS Extensometer 1 Volt = 3.96x10 strain (in./in.)

2.5.3 Tensiletestresults
15B

Two distinct steel stressstrain curves were observed in this study. Tensile
coupons from the 362S16233 structural studs demonstrate gradual yielding behavior,
while the tensile coupons from the 600S16233 studs demonstrated a sharp yielding
plateau.Theyieldstress,Fy,forthegraduallyyieldingspecimenswasdeterminedwith
the 0.2% strain offset method. The stressstrain curve for specimen 362348NH (East
Flange)demonstratestheoffsetmethodin Figure5.30.Theyieldstressforthesharply
139H

yielding specimens was determined by averaging the stresses in the yield plateau.
ASTMdoesnotprovidespecificguidelinesonhowtoaveragetheplateaustresses.For
this autographic method, the averaging range is determined by using two strain offset
lines,oneat0.4%strainoffsetandtheotherat0.8%offsetasshownforspecimen60024

194

NH (West Flange) in Figure 5.31. The steel modulus of elasticity, E, was assumed as
1394H

29500ksiforallspecimenswhendeterminingtheyieldstress.Thetensilecouponyield
stresses and cross section dimensions are summarized in Table 5.13. The mean and
1395H

standard deviation for all 362S16233 and 600S16233 tensile coupons tested are
providedinTable5.14.
1396H

100

0.2% strain offset line (slope=29500 ksi)

90

Axial Tensile Stress (ksi)

80

0.8

70
0.6

60
50

0.4

40
30

YIELD STRESS (0.2% offset)=60.1 ksi

0.2

20
0

10
0

0.2
0.05

0.4

0.6

0.1
0.15
Engineering Strain,(in./in.)

0.8
0.2

1
0.25

Figure5.30Graduallyyieldingstressstraincurvewith0.2%strainoffsetmethod
100

0.4% strain offset line (slope=29500 ksi)

90

0.8% strain offset line (slope=29500 ksi)

Axial Tensile Stress (ksi)

80

0.8

70
0.6

60
50

0.4

40
30

YIELD STRESS (Autographic Method)=59.7 ksi

0.2

20
0

10
0

0.2
0.05

0.4

0.6

0.1
0.15
Engineering Strain (in./in.)

0.8
0.2

1
0.25

Figure5.31SharpyieldingstressstraincurveusinganautographicmethodfordeterminingFy
Table5.13Summaryofcolumnspecimensteelyieldstress

195

Specimen

tbase,w
in.

Web
wmin
in.

Fy
ksi

362-1-24-NH
0.4985
53.3
362-2-24-NH 0.0368
362-3-24-NH
362-1-24-H
0.0390
0.4945
55.9
362-2-24-H
0.0368
0.4886
52.9
362-3-24-H
0.0394
0.4945
55.6
362-1-48-NH 0.0392
0.4985
59.4
362-2-48-NH 0.0393
0.4990
59.2
362-3-48-NH 0.0389
0.4930
58.0
362-1-48-H
0.0391
0.4998
59.5
362-2-48-H
0.0390
0.4992
58.8
362-3-48-H
0.0401
0.4990
57.8
600-1-24-NH
0.4950
60.6
600-2-24-NH 0.0438
600-3-24-NH
600-1-24-H
0.0414
0.4899
61.9
600-2-24-H
0.0427
0.4964
57.8
600-3-24-H
0.0429
0.4966
59.7
600-1-48-NH 0.0434
0.4985
58.7
600-2-48-NH 0.0435
0.4985
N/C
600-3-48-NH 0.0436
0.4995
60.4
600-1-48-H
0.0429
0.4970
60.3
600-2-48-H
0.0429
0.4994
61.8
600-3-48-H
0.0430
0.4992
60.7
NOTE: N/C Tests results were not obtained

tbase,f1
in.

West Flange
wmin
in.

Fy
ksi

tbase,f2
in.

East Flange
wmin
in.

Fy
ksi

0.0415

0.4975

54.7

0.0372

0.4955

57.4

0.0391
0.0390
0.0394
0.0393
0.0394
0.0391
0.0393
0.0391
0.0400

0.4963
0.4950
0.4927
0.4965
0.4975
0.5000
0.4985
0.4961
0.4957

59.3
58.8
N/C
59.7
59.3
58.9
58.2
60.6
58.0

0.0391
0.0391
0.0394
0.0392
0.0393
0.0390
0.0394
0.0391
0.0397

0.4968
0.4945
0.4947
0.4975
0.4970
0.4930
0.4991
0.4975
0.4978

58.5
59.5
56.4
59.9
59.2
60.1
58.1
59.8
59.1

0.0432

0.4950

59.7

0.0438

0.5000

55.9

0.0422
0.0384
0.0431
0.0436
0.0430
0.0432
0.0426
0.0428
0.0434

0.4940
0.4874
0.4954
0.4955
0.4970
0.4955
0.4980
0.4962
0.4961

63.6
55.6
58.0
62.3
63.4
N/C
63.0
62.1
59.7

0.0428
0.0424
0.0430
0.0434
0.0430
0.0433
0.0429
0.0431
0.0430

0.4964
0.4938
0.4960
0.4965
0.4970
0.4965
0.4970
0.4977
0.4977

60.3
61.8
62.6
59.3
63.3
61.9
60.8
62.2
64.0

Table5.14Columnspecimensteelyieldstressstatistics
Stud Type
362S162-33
600S162-33

yield stress, Fy
mean
STDV
ksi
ksi
58.1
2.0
61.0
2.0

5.3 Elasticbucklingcalculations
36B

Elastic buckling provides a means to categorize and potentially better understand

theloaddeformationresponseandultimatestrengthofthethinwalledcolumnsinthis
study. The local, distortional, and global elastic buckling modes and their associated
critical elastic buckling loads (Pcrl, Pcrd, Pcre) are presented here for each specimen.
Calculations are performed with a shell finite element eigenbuckling analysis as
opposedtoananalysisusingFSM(Schaferanddny2006)tocapturetheinfluenceof
theslottedwebholesandthetested(fixedfixed)boundaryconditions.

196

3.1 Finite element modeling assumptions


81B

Eigenbuckling analysis in ABAQUS is performed for the 24 column specimens

(ABAQUS 2007a). All columns are modeled with S9R5 reduced integration ninenode
thin shell elements. Coldformed steel material properties are assumed as E=29500 ksi
and =0.30. The centerline Csection dimensions input into ABAQUS are calculated
using the outtoout dimensions and flange and lip angles at the midheight of each
column specimen as provided in Table 5.3 and Table 5.4. Each column specimen is
1397H

1398H

loadedwithasetofconsistentnodalloadsinABAQUStosimulateaconstantpressure
across the bearing edge of the specimen. The nodes on the loaded column face are
coupled together in the direction of loading with an ABAQUS pinned rigid body
constraint(seeFigure4.12).
139H

3.2 Elastic buckling results


82B

3.2.1 Buckledshapes/eigenmodes
152B

The first (lowest buckling load) local (L) and distortional (D) buckled shapes for

specimenswithandwithoutslottedholesarecomparedin Figure5.32and Figure5.33.


140H

140H

TheLandDmodesforeachspecimenwereidentifiedvisuallybymanuallysearching
throughtheelasticbucklingmodesproducedintheeigenbuckinganalysis.Thenominal
crosssection halfwavelengths in Table 5.1 were compared to the halfwavelengths in
1402H

thefiniteelementmodeltoassistinthecategorization.Thelocalanddistortionalmodes
thatmostresembledtheFSMresultsforLandDmodeswereselected.Thismethodof
modal identification is neither exact nor ideal, especially when both local and

197

distortional buckling are present in the same eigenmode. Formal modal identification
hasrecentlybeendevelopedinthecontextofthefinitestripmethod(Schaferanddny
2006)andfutureworkisongoingtoextendthismethodtofiniteelementanalysesandto
problemssuchastheonesencounteredhere.
(a)

(b)

Holes cause
mixed
distortional-local
mode

Hole
terminates
web local
buckling

Hole
terminates
web local
buckling

Local Buckling (L)

Distortional Buckling (D)

Local Buckling

Distortional Buckling

Figure5.32(a)Localanddistortionalelasticbuckledmodeshapesfor(a)short(L=48in.)362S16233
specimensand(b)intermediatelength(L=48in.)362S16233specimens.

(a)

(b)

Holes cause mixed


distortional local
mode

Local Buckling (L)

Hole changes
number of
half-waves
from 5 (NH)
to 6 (H)

Distortional Buckling (D)

Local Buckling

Holes
change
number of
half-waves
from 8 (NH)
to 12 (H)

Distortional Buckling

Figure5.33Localanddistortionalelasticbuckledmodeshapesfor(a)short(L=48in.)600S16233specimens
and(b)intermediatelength(L=48in.)600S16233specimens.

3.2.2 Bucklingloads/eigenvalues
153B

TheprimarygoalofthisresearchprogramistoextendtheDirectStrengthMethodto

coldformedsteelstructuralmemberswithholes.TheDirectStrengthMethod(DSM),a
design method for coldformed steel structural members, predicts column ultimate
strength by predicting the column failure mode and ultimate strength through

198

knowledge of the local (L), distortional (D), or global (G) elastic buckling modes. This
connection is made using the critical elastic buckling load, Pcr, and the slenderness,
definedwiththeratioofcolumnsquashloadPygtoPcrfortheL,D,andGmodes. Table
1403H

5.15summarizesPcrandPygforthespecimensevaluatedinthisstudy.Thesquashload
Pyg is calculated with the gross crosssectional area, and Pcr includes the effects of the
holesandthetested(fixedfixed)boundaryconditions.(Note,theimplicationsofusing
PygasopposedtoPy,netatthenetsectionarediscussedinChapter8.)
140H

The influence of holes on Pcr is of interest in the context of DSM because elastic

buckling loads and slenderness are used to predict ultimate strength. To isolate the
influenceofholesonPcr,additionaleigenbucklinganalysesofthespecimenswithholes
(specimens labeled with an H) were performed, but with the holes removed (the
boundaryandloadingconditionswerenotmodifiedandthemeshusedinthemodels
wasidenticalexceptfortheremovedelementsattheholelocation).Thecomparisonof
Pcrforspecimenswithholes(H)andthenwithholesremoved(noH)isalsosummarized
inTable5.15.
1405H

199

Table5.15Criticalelasticbucklingloads,influenceofholesonelasticbuckling
Specimen
Name

Pyg
kips

ELASTIC BUCKLING
Pcrl
Pcre
kips

HOLE INFLUENCE*
Pcrd

kips

Pcre/Pcre

kips

noH

noH

Pcrl/Pcrl

Pcrd/Pcrd

362-1-24-NH
15.5
109.4
4.9
10.6
N/A
362-2-24-NH
15.6
112.5
4.8
10.2
362-3-24-NH
15.7
112.2
5.0
10.7
362-1-24-H
16.4
119.3
5.9
13.5
0.98
1.03
1.12
362-2-24-H
15.7
112.8
5.4
12.4
0.98
1.02
1.13
362-3-24-H
16.4
130.6
5.7
12.9
0.99
1.02
1.12
362-1-48-NH
16.9
30.5
5.2
9.7
N/A
362-2-48-NH
16.7
29.5
5.2
9.6
362-3-48-NH
16.6
29.6
5.1
9.5
362-1-48-H
16.6
30.0
5.3
9.4
0.94
1.03
0.98
362-2-48-H
16.8
29.7
5.2
9.3
0.94
1.03
0.98
362-3-48-H
16.8
36.2
5.7
9.6
0.95
1.03
0.98
600-1-24-NH
24.7
244.5
3.4
6.8
N/A
600-2-24-NH
24.5
234.9
3.4
6.7
600-3-24-NH
24.5
218.4
3.4
6.6
600-1-24-H
25.0
239.3
3.3
7.0
1.01
1.02
1.09
600-2-24-H
23.1
238.4
3.2
6.7
1.01
1.01
1.08
600-3-24-H
24.7
242.6
3.5
7.3
1.02
1.01
1.08
600-1-48-NH
25.1
61.8
3.5
5.2
N/A
600-2-48-NH
26.2
59.6
3.4
5.7
600-3-48-NH
25.4
60.2
3.4
5.7
600-1-48-H
25.2
56.3
3.4
5.1
0.87
1.02
1.02
600-2-48-H
25.5
53.0
3.4
5.0
0.87
1.02
1.02
600-3-48-H
25.6
55.8
3.4
5.0
0.86
1.02
1.02
* For specimens with holes (H), the holes are removed and elastic buckling calculated (noH).
The hole (H) and no hole (noH) finite element models are otherwise identical, isolating the influence of the holes.

noH

3.2.3 Modalinteractionatultimatestrength
154B

An additional reason for the selection of these specimen crosssections, at these

lengths, beyond the reasons discussed in Section 5.2.1, is that the specimens provide
1406H

much needed experimental data on crosssections with potential modal interaction at


ultimate strength both with and without holes. Typically modal interaction is
understoodtobeaconcernwhentheelasticbucklingloadsofmultiplemodesareator
near the same value, and the ratio of any two elastic bucking loads (e.g., Pcrl/Pcrd) is
considered a useful parameter for study. However, for modes with different post
buckling strength and where material yielding is considered, a more pressing concern
may be the situation when both failure modes predict similar capacities. Which mode
doesthecolumnfailinifthepredictedcapacityinlocal(Pnl)anddistortional(Pnd)areat
or near the same level? What impact does a hole have on the failure mode that is

200

triggered?Inthespecimensselectedhere,theratioofPcrl/Pcrdvariesfromaminof0.44to
amaxof0.68,butisnevernear1.0.Therefore,bythistraditionalmeasurenomeaningful
interactionwouldbeanticipated.However,iftheDSMmethodologyisusedtopredict
thecapacities,asillustratedin Figure5.34,thepredictionsfortheratioofthetwolimit
1407H

statesPnl/Pndrangesfromaminof0.86toamaxof0.90inthe362S16233shortcolumns
and from a min of 1.0 to a max of 1.05 in the600S16233 short columns (the ratios are
similarforthelongcolumnspecimens).Thus,thesecrosssectionsprovideameansto
examine the potential for localdistortional modal interaction at ultimate strength, and
offer valuable data for determining any necessary modification to the DSM
methodologywhenholesarepresent.

1.2

Local DSM
Distortional DSM

Pn/Pyg

0.8

362S162-33 short columns

0.6

600S162-33 short
columns

L
D

0.4
Local(L)-Distortional(D) interaction is
expected since predicted strengths (Pn)
are of similar magnitudes

0.2

0.5

1
1.5
2
slenderness, or d
l

2.5

Figure5.34Local(L)anddistortional(D)DSMstrengthpredictionsaresimilarinmagnitudeforboth
362S16233and600S16233crosssections,indicatingthatLDmodalinteractionwilloccurduringthetested
responseofthecolumns.

201

3.3 Discussion of elastic buckling results


83B

3.3.1 Localbuckling
15B

Boundary conditions have little influence on the local buckling mode shapes

(comparedwithFSMLmodes),butthepresenceoftheslottedwebholescanchangethe
shape, halfwavelength, and buckling load of the first (lowest) local buckling mode
observed. In the 362S16233 specimens the web holes terminate local buckling in the
vicinityoftheholes,see Figure5.32.Inthe600S16233specimensthewebholescause
1408H

anincreasednumberofhalfwavesalongthelengthtooccurinthelowestlocalmode,
see Figure 5.33. The presence of holes causes a slight increase in Pcrl (see Table 5.15)
1409H

140H

which is consistent with the increased number of observed local buckling halfwaves.
Themoreextensiveelasticbucklingstudies Chapter3and Chapter4demonstratethata
14H

142H

holecanincreaseordecreasethenumberofbuckledhalfwaves(andthecriticalelastic
bucklingload)ofrectangularplatesandcoldformedsteelstructuralstuds.

3.3.2 Distortionalbuckling
156B

Boundary conditions and the presence of holes have an influence on the observed

distortionalbucklingmodeshapes(comparedwithFSMDmodes)andbucklingloads.
The boundary conditions (fixedfixed) allow a smaller number of halfwaves to form
than predicted using the simply supported FSM D modes of Table 5.1. For example,
143H

observe the restrained shape of the buckled distortional halfwave near the member
ends in Figure 5.32a. In longer specimens (see Figure 5.32b and Figure 5.33b), the
14H

145H

146H

influence of the boundary conditions lessens and the halfwavelength of distortional

202

bucklingatmidheightapproachesthatoftheFSMDmodeofTable5.1.(Section4.2.6.2
147H

148H

explores the influence of fixedfixed boundary conditions on Pcrd using the column
experiment database.) The presence of the web holes complicates the predicted D
modes,seeFigure5.32andFigure5.33.LocalbucklingnowappearswithintheDmode
149H

1420H

itself.ThehalfwavelengthoftheseinteractingLmodesissignificantlyshorterthanthe
lowest L modes observed. Further, and rather unintuitively, the buckling load, Pcrd,
actuallyincreaseswiththepresenceofholesintheshortcolumnspecimens(asmuchas
13%).However,thisincreaseislostatthelongerspecimenlengthwherethemaximum
changeinthebucklingloadis+/2%.Thisresultsuggeststhatintheshorterspecimens
theremovalofthematerialmostsusceptibletooutofplanebending,atthemiddepth
of the web, actually serves to stiffen the column (a localized increase in the transverse
bendingstiffnessofplateswithholeshasbeenobserved,seeFigure4.30).Thisinfluence
142H

does not persist in the longer specimens suggesting that the increased stiffness is only
relevantwhentheDmodeisatarestrainedhalfwavelength.Thus,iftheDmodeisfree
to form (over a long enough unbraced length) the holes do not increase the elastic
bucklingload.

3.3.3 Globalbuckling
157B

The global (Euler) buckled shapes for the intermediate 362S16233 and 600S16233

columns in Figure 5.35 occur as flexuraltorsional buckling, although local and


142H

distortional deformation are both present in the mode shape for specimens with and
withoutholes,whichisanunexpectedresult.Theinteractionbetweentheglobal,local,

203

anddistortionalmodesmakestheidentificationoftheglobalmodedifficult.TheEuler
bucklingloadandmodeshapepredictedwithclassicalmethods(inCUTWP),whichdo
notallowcrosssectiondistortionandignoreholes,wereusedtodeterminetherangeof
bucklingloads(eigenvalues)tobevisuallysearched.ThereportedmodesinFigure5.35
1423H

are the ones closest to the expected buckling load exhibiting significant global
deformations. Additional eigenbuckling analyses of the 362S16233 and 600S16233
crosssectionswereperformedatalongercolumnlength(8ft.)andtheseanalysesshow
no local or distortional interaction with the global modes. Therefore, the observed
interaction is length dependent and not a fundamental feature of global buckling in
thesecrosssections.AnalternativehypothesisfortheunusualmodeshapesinFigure
142H

5.35 is that several buckling mode shapes exist near the global critical elastic buckling
load,whichcausestheeigensolvertomisreporttheglobalmodeasalinearcombination
ofbuckledshapes.

As for the global buckling loads, the slotted holes have a small influence on the

globalbucklingloadfortheintermediatelength362S16233specimens,reducingPcrebya
maximum of 6%. However, Pcre for the intermediate length 600S16233 columns
decreasesbyamaximumof14%withthepresenceofthetwoslottedholes,whichisan
unexpected result attributed to the local and distortional modes mixing with global
buckling (i.e., Figure 5.35). Additional research work is ongoing to determine under
1425H

whatconditionsholesinfluencetheglobalcriticalelasticbucklingload.

204

Global (Euler) Buckling


Finite element eigenbuckling
analyses predict global
buckling interacting with local
and distortional buckling
362-1-48-NH

600-1-48-NH

362-1-48-H

600-1-48-H

CUTWP predictions using


classical stability theory

Figure5.35Comparisonofglobalmodeshapesforintermediatelength362S16233and600S16233
specimens.

5.4 Experimentresults
37B

4.1 Ultimate strength


84B

The peak tested compressive load for all column specimens and an average peak

loadforeachtestgroupareprovidedinTable5.16.Theslottedholesareshowntohave
1426H

onlyasmallinfluenceoncompressivestrengthinthisstudy,withthelargestreduction
being2.7%forthe362S16233shortcolumns.

205


Table5.16Specimenultimatestrengthresults
Specimen
362-1-24-NH
362-2-24-NH
362-3-24-NH
362-1-24-H
362-2-24-H
362-3-24-H
362-1-48-NH
362-2-48-NH
362-3-48-NH
362-1-48-H
362-2-48-H
362-3-48-H
600-1-24-NH
600-2-24-NH
600-3-24-NH
600-1-24-H
600-2-24-H
600-3-24-H
600-1-48-NH
600-2-48-NH
600-3-48-NH
600-1-48-H
600-2-48-H
600-3-48-H

Ptest
kips
10.48
10.51
10.15
10.00
10.38
9.94
9.09
9.49
9.48
8.95
9.18
9.37
11.93
11.95
12.24
12.14
11.62
11.79
11.15
11.44
11.29
11.16
11.70
11.16

Mean
kips

Std. Dev.
kips

10.4

0.2

10.1

0.2

9.4

0.2

9.2

0.2

12.0

0.2

11.9

0.3

11.3

0.1

11.3

0.3

4.2 Failure modes and post-peak ductility


85B

4.2.1 Shortcolumns
158B

Theloadingprogressionforthe362162S33shortcolumnsisdepictedin Figure5.36
1427H

(withoutahole)andFigure5.37(withahole).Bothcolumnsexhibitlocalbucklingofthe
1428H

webnearthesupportscombinedwithonedistortionalhalfwavealongthelength.This
distortional buckling pattern is consistent with that predicted by the elastic buckling
modeshapesof Figure5.32a.Forthecolumnwiththehole,localizedholedeformation
1429H

(Figure5.37,rightmostpicture)initiatesataloadofapproximately0.4Ptestandincreases
1430H

in magnitude as the test progresses. This observed deformation behavior is visually


consistent with the unstiffened strip approach discussed in Error! Reference source

206

notfound.,wherethestripofweboneithersideoftheholeisassumedtobehaveasan
unstiffenedelement.

Theinwardflangedeformationconcentratesattheholeafterpeakloadintheshort

362S16233 specimens with holes. It is hypothesized that the slotted hole reduces the
postpeakresistanceoftheweb,causingtheflangesandlipstocarrymoreofthecolumn
load.Thisreductioninpostpeakresistanceisquantifiedbyobservingthereductionin
areaundertheloaddisplacementcurveforthecolumnwiththeslottedhole,asshown
inFigure5.38.
143H

Distortional buckling
in one half-wave at
peak load

(a) P=0 kips

(b) P=7.0 kips

(c) P=10.5 kips


(peak load)

(d) P=7.5 kips

Figure5.36Loaddisplacementprogressionforshortcolumnspecimen362224NH

207

Local buckling at hole


(unstiffened strip)

(a) P=0 kips

(b) P=10.4 kips


(peak load)

(c) P=7.0 kips

Figure5.37Loaddisplacementprogressionforshortcolumnspecimen362224H

14

Slotted hole has small


influence on peak load

362-2-24-NH
362-2-24-H

Column axial load (kips)

12
10
8
6
4

Slotted hole influences post-peak


load path and reduces column
ductility

2
0

0.05
0.1
0.15
Column axial displacement (inches)

0.2

Figure5.38Loaddisplacementcurvefora362S16233shortcolumnwith,withoutaslottedhole

208

Positiontransducersplacedatthemidheightoftheshortcolumnspecimenscapture

therateoflateralflangedisplacementassociatedwithdistortionalbuckling,D,asshown
in Figure 5.39. Figure 5.39 demonstrates that the initiation of web local buckling does
1432H

143H

notinfluencetheaxialstiffnessofspecimen362224NH,butratherthatasofteningof
the loadaxial deformation curve coincides with the increased rate of lateral flange
movement (distortional buckling). This observation suggests that the loss in axial
stiffnessassociatedwithdistortionalbucklingplaysalargerrolethanweblocalbuckling
inthepeakloadresponseofthe362S16233shortcolumns.Theinfluenceoftheslotted
hole on lateral flange displacement is provided in Figure 5.40, where the postpeak
143H

flangedisplacementratesaresignificantlyhigherforthe362S16233shortcolumnwith
holes.Theresultsof Figure5.40indicatethatholespotentiallyhaveasignificantimpact
1435H

on the collapse mechanisms triggered from distortional buckling. Lateral flange


displacementplotsareprovidedforallspecimensinAppendixF.
1436H

209

12

2
Local buckling
half-waves
first observed

1.5
8
Rate of flange distortion
increases as loaddisplacement curve softens

D (inches)

Column axial load (kips)

10

4
0.5
2

D =
+ west

west + east

+ east

2
0
0.2

0.05
0.1
0.15
Column axial displacement (inches)

Figure5.39Comparisonofloaddeformationresponseandlateralflangedisplacementsforspecimen3622
24NH

2
362-2-24-NH
362-2-24-H

1.8
1.6

Increased rate of flange distortion


is observed for column with a hole
after peak load is reached

D (inches)

1.4
1.2
1
H

0.8
Peak load
occurs here

0.6

NH

0.4

D =

0.2
0

+ west

+ east

0.05
0.1
0.15
Column axial displacement (inches)

west + east
2
0.2

Figure5.40Influenceofaslottedholeon362S16233shortcolumnlateralflangedisplacement

210

Figure5.41and Figure5.42depictthedeformationresponseofthe600S16233short
1437H

1438H

columns with and without a slotted hole. In both cases, local buckling at the loaded
ends combines with one distortional halfwave along the column length. The
distortionalbucklingpatternforthesespecimensisnotwhollyconsistentwiththeelastic
bucklingpredictionsofFigure5.33a,whichshowstwodistortionalhalfwaves;however,
1439H

specimens600224Hand600324Hdidbuckleintwohalfwaves,seeAppendixFfor
140H

pictures.Theseresultssuggestthatgeometricimperfectionsalsohavearoletoplayin
the details of the buckling mode initiated in the loaded response. The deformation
response of the member with and without the hole is similar through the test
progression,suggestingthattheholehasasmallinfluenceoncompressivestrengthand
postpeakductilityfortheholewidthtowebwidthratiosconsideredhere. Figure5.43
14H

confirmsthattheslottedholehasaminimaleffectonthepostpeakloadresponseofthe
column.

211

Distortional
buckling in one
half-wavelength
at peak load

Web local
buckling

(a) P=0 kips

(b) P=8.0 kips

(c) P=11.9 kips


(peak load)

(d) P=8.0 kips

Figure5.41Loaddisplacementprogressionforshortcolumnspecimen600124NH

Web local
buckling
initiates

(a) P=0 kips

Local buckling near


supports combines
with distortional
buckling (one halfwave) at peak load

(b) P=7.5 kips

(c) P=12.1 kips


(peak load)

Similar failure
mode to no hole
specimens

(d) P=8.0 kips

Figure5.42Loaddisplacementprogressionforshortcolumnspecimen600124H

212

14
600-1-24-NH
600-1-24-H

Column axial load (kips)

12
10
8
6
NH

Slotted hole has small


influence on post-peak
response and ductility

2
0

0.05
0.1
0.15
Column axial displacement (inches)

0.2

Figure5.43Comparisonofloaddisplacementresponseforshort600S16233columnspecimenswithand
withoutholes

4.2.2 Intermediatelengthcolumns
159B

Figure5.44and Figure5.45summarizethedeformationresponseofthe362S16233
142H

143H

intermediatelengthcolumnswithandwithoutholes.Inbothcases,localwebbuckling
is first observed at approximately 0.45Ptest which is lower than, but the same order of
magnitudeas,thecalculatedlocalcriticalelasticbucklingloadPcrl.Notein Figure5.45
14H

thatthelocalbucklinghalfwavesaredampenedinthevicinityoftheholes,similarto
the elastic buckling prediction of Figure 5.32b. Further, the observed local buckling
145H

wavesareathalfwavelengthsconsistentwiththelocalbucklingmodein Figure5.32b,
146H

notthoseshowninteractingwithdistortionalbuckling.(Thisobservationsupportsthe
ideathatthefundamentalelasticbucklingmodesL,D,andGarerepresentativeofthe
physicalbehaviorofthecolumnandthatthemixedmodesobservedinaneigenbuckling
analysis only exist numerically.) Three distortional buckling halfwaves become well

213

formedatapproximately0.70Ptest,overcomingthelocalhalfwavesinthewebexceptat
themidheightofthecolumn.Thisdistortionalbucklingpatternisconsistentwiththe
elasticbucklingpredictioninFigure5.32b.Figure5.46demonstratesthatthepresenceof
147H

148H

slottedholeshasonlyaminimalinfluenceonloadaxialdisplacementresponse.

All of the 362S16233 intermediate length columns failed soon after the peak load

with a sudden loss in loadcarrying capacity caused by global flexuraltorsional


buckling. Yielding of the column flanges reduces the torsional stiffness of the section,
and the friction end conditions could not restrain the twisting of the column. The
twistingofspecimen362348NHisquantifiedin Figure5.47asthedifferencebetween
149H

the west and east midheight flange displacements, T, captured by the position
transducers.Thelateraldisplacementoftheflangetipsduetodistortionalbuckling(D),
also shown in Figure 5.47, is separated from the twisting effect by averaging the west
1450H

and east midheight flange displacements. Figure 5.47 shows that the crosssection is
145H

bothopeningandtwisting,butitistheabruptincreasein Toccurringwellpastpeak
loadthatleadstothecollapseofthememberinthetest.

214

Web local buckling


with flange distortion
in three half-waves

Local buckling mixes


with distortional halfwave at peak load

Pure distortional
buckling dominates over
local buckling in this
half-wave

(a) P=0 kips

(b) P=8.0 kips

(c) P=9.5 kips


(peak load)

Figure5.44Loaddisplacementprogression,intermediatelengthcolumnspecimen362348NH

Hole dampens web local buckling by


creating two stiff web strips on either side of
hole that locally boost Pcrl above Pcrd

(a) P=0 kips

(b) P=8.0 kips

(c) P=9.4 kips


(peak load)

Figure5.45Loaddisplacementprogressionforintermediatelengthcolumnspecimen362348H

215

14

Columns fail abruptly


with a global torsional
buckling mode

Column axial load (kips)

12

362-3-48-NH
362-3-48-H

10
8
6
4

NH

2
H

0.05
0.1
0.15
Column axial displacement (inches)

0.2

Figure5.46Loaddisplacementcurve,362S16233intermediatecolumnwithandwithoutahole

2
1.5
Abrupt failure

D, T (inches)

1
0.5

T =

east west
2

-0.5
-1
+ east

+ west

-1.5
-2

Peak load
occurs here

D =

east + west

0.05
0.1
0.15
Column axial displacement (inches)

2
0.2

Figure5.47362S16233longcolumnmidheightflangedisplacementsshowtheglobaltorsionalfailure
mode

216

The loaddisplacement response for the intermediate length 600S16233 columns

with and without slotted holes is depicted in Figure 5.48 and Figure 5.49. Local
1452H

1453H

bucklingisobservedatapproximately0.45Ptestforbothsections.Theholesdonotrestrict
the local buckling halfwaves as was the case in the 362S16233 intermediate length
columns. This local buckling behavior is consistent with that observed in the elastic
bucklinganalysis,see Figure5.33b.Threedistortionalhalfwavesformasthecolumns
145H

(all3ofthe600S16233intermediatelengthspecimens)approachpeakload.Twoloud
soundsresonatefromthecolumnsnearpeakloadasthelocalwebbucklinghalfwaves
at the two column ends abruptly snap into one distortional halfwave per end. The
changefromlocaldominatedtodistortionaldominatedwebbucklingisreflectedastwo
drops in the loaddisplacement response near peak load for the 600S16233 column
withoutholes,asshowninFigure5.50.The600S16233columnwithslottedholesisnot
145H

affectedbythisabruptmodeswitching,asitmaintainsweblocalbucklingwellbeyond
peak load. The observations suggest that in this case the holes are beneficial because
theymaintainthelocalbucklinghalfwavesthroughpeakload,allowingthecolumnto
relymoreonthepostpeakstrengthprovidedbythebuckledweb.Thismodeswitching
is a difficult challenge for numerical models and these results, repeated in 3 tests,
provides an important and challenging experimental benchmark for the numerical
modelingofthesemembersSection7.2.
1456H

217

Multiple local half-waves change to one


distortional half-wave with loud resonant
sound at peak load

(a) P=6.0 kips

(b) P=11.2 kips (c) P=11.2 kips (d) P=7.0 kips


(peak load)

Figure5.48Loaddisplacementprogression,intermediatelengthcolumnspecimen600148NH

Hole preserves web local


buckling through peak load

(a) P=0 kips

(b) P=11.2 kips


(peak load)

(c) P=10.0 kips

Figure5.49Loaddisplacementprogression,intermediatelengthcolumnspecimen600148NH

218

14
600-1-48-NH
600-1-48-H

Column axial load (kips)

12
10
8

Drops in load occur when


multiple web local buckling
half-waves change abruptly to
one distortional half-wave
(north end first, then south end)

6
4
2
0

0.05
0.1
0.15
Column axial displacement (inches)

0.2

Figure5.50Loaddisplacementcomparisonofintermediatelength600S16233specimenswithandwithout
holes

4.3 Discussion of hole influence on elastic buckling and tested response


86B

Bothlocalanddistortionalelasticbucklingwereobservedinthetestedresponseof

the specimens and contributed in different ways to the failure modes of the columns.
Local buckling initiated plastic folding in the web at peak load, and distortional
buckling was reflected as either opening (D) or closing (+D) of the crosssection and
yieldingoftheflangesandlipstiffeners.Allthreeoftheshort362S16233columnswith
holes exhibited a closed distortional buckling failure (+D), where the presence of the
slottedholeconcentratedtheplasticdeformationintheflangesandlipsadjacenttothe
hole.Thisresultwasdifferentfromtheshort362S16233columnswithoutholeswhere
mixed localdistortional failures were observed. The slotted holes also changed the

219

bucklinginfluenceatpeakloadintheintermediatelength600S16233specimens,where
the holes prevented local web buckling from switching to distortional buckling in all
threespecimentests.Thedeformationatpeakloadfortheintermediatelength362S162
33 and short 600S16233 specimens was less sensitive to the presence of slotted holes,
exhibiting mixed localdistortional failure modes consistent with DSM predictions (L
andDofsimilarmagnitudes)asdiscussedinSection5.3.2.3.
1457H

The visual observations in this study highlight the complex relationship between

elasticbucklingandcolumnfailureandthesensitivityoftheirinteractiontothechoice
of crosssection and column length. In the cases of the short 362S16233 and
intermediate length 600S16233 specimens, it is clearly demonstrated that holes can
influence column deformation and ductility by changing how elastic buckling modes,
localanddistortionalinthiscase,affecttheaxialstiffnessandplasticdeformationofthe
columnunderload.ThisdataisimportantinthecontextoftheDirectStrengthMethod,
especially for this current effort to extend DSM to members with holes, since elastic
bucklingisusedtopredictthefailuremode(local,distortional,orglobal)andultimate
strength.

4.4 Discussion of friction-bearing boundary conditions


87B

The frictionbearing end conditions used in this testing are advantageous because

specimen alignment and preparation can be performed without welding or the use of
grout or hydrostone. The specimens were aligned by hand in the testing machine
without special equipment. However, preparing the specimen ends with a milling

220

machinecanbetimeconsuming.Further,smalldeviationsinflatnessmaysignificantly
impactthetestedresultsandfailuremodes;realcaremustbetakeninthespecimenend
preparation.Finally,lackofapositiveconnectionbetweenspecimenandplatenmakes
itdifficulttoexactlyknowtheboundaryconditions.

Inthisstudy,frictionbetweenthecolumnendsandtheplatenspreventedachange

inshapeofthecrosssectionuptopeakloadinallspecimens,butslippingofthecross
section was observed after peak load. This slipping was signaled by loud metalon
metalpoppingsoundsassociatedwithobservablechangesinthecrosssection(Dof
the flanges, see Figure 5.47 for definition) at the column ends. Also, uplift warping
1458H

deformations like those shown in Figure 5.51 occurred in the postpeak range for the
1459H

short600S16233columnsexperiencingdistortionaltypefailures.Distortionalbuckling
modes are anticipated to be sensitive to this uplift since they are highly sensitive to
warping deformations. The intermediate length 362S16233 columns experienced a
suddenglobalflexuraltorsionalfailureshortlyafterreachingpeakloadasthetwisting
of the columns overcame the friction between the column ends and the platens. The
frictionbearing end conditions did not allow a detailed study of the global flexural
torsionalpostpeakresponsefortheintermediatelength362S16233columnsandlikely
decreasedtheirultimatestrengths.

Overall, for short and intermediate length column testing focused on local and

distortional buckling modes, the advantages of the simple frictionbearing boundary


conditionsoutweighedthedisadvantages.Propercaremustbetakentoinsuretheends
are milled flat and the platens are level and parallel. For longer column tests, where

221

large torsional rotations must be restrained, the bearing conditions employed here are
notrecommendedforuse.

Flange-lip corner lifts off platen


when large deformations exist
past peak load

Figure5.51Short600S16233columnflangelipcornerliftsoffplatenduringpostpeakportionoftest

222

Chapter 6
Predicting residual stresses and plastic
strains in cold-formed steel members
5B

Thincoldformedsteelmembersbeginasthick,molten,hotsteelslabs.Eachslabis

typically hotrolled, coldreduced, and annealed before coiling and shipping the thin
steel sheet to rollforming producers (US Steel 1985). Once at a plant, the sheet is
unwoundthroughaproductionlineandplasticallyfoldedtoformthefinalshapeofa
structuralmember,asshowninFigure6.1.Thismanufacturingprocessimpartsresidual
1460H

stresses and plastic strains through the sheet thickness. These residual stresses and
strains influence the loaddisplacement response and ultimate strength of coldformed
steelmembers.

In previous work a statistical approach was employed to draw conclusions on the

magnitudeanddistributionoflongitudinalresidualstressesusingadatasetofsurface
strainmeasurementscollectedbyresearchersbetween1975and1997(SchaferandPekz
1998). The measured surface strains are converted to residual stresses using Hookes

223

Law and then distributed through the thickness as membrane (constant) and bending
(linearvariation)components.Theseresidualstressdistributionsareaconvenientway
to express the measured residual surface strains, and areconvenient as well for use in
nonlinear finite element analyses, but they are not necessarily consistent with the
underlyingmechanics.

Figure6.1Coldformedsteelrollforming:(left)Sheetcoilentersrollformingline,(right)steelsheetiscold
formedintoCshapecrosssection(photoscourtesyofBradburyGroup).

Plastic bending, followed by elastic springback, creates a nonlinear through

thicknessresidualstressdistribution,inthedirectionofbending,asshownin Figure6.2
146H

(Shanley 1957). The presence of nonlinear residual stress distributions in coldformed


steel members has been confirmed in experiments (Key and Hancock 1993) and in
nonlinearfiniteelementmodelingofpressbrakingsteelsheets(Quachetal.2006).A
closedform analytical prediction method for residual stresses and equivalent plastic
strains from coiling, uncoiling, and mechanical flattening of sheet steel has also been
proposed (Quach et al. 2004 ). The same plastic bending that creates these residual
stressesalsoinitiatesthecoldworkofformingeffect,whereplasticstrainsincreasethe
apparentyieldstressinthesteelsheet(andultimatestrengthinsomecases)(Yu2000).

224

Together,theseresidualstressesandplasticstrainscomprisetheinitialmaterialstateof
acoldformedsteelmember.
Elastic springback

Plastic bending

Elastic springback

Plastic bending

compression

+
tension

Plastic bending

Elastic springback

=
Nonlinear residual
stress distribution

Figure6.2Formingabend:plasticbendingandelasticspringbackofthinsheetsresultsinanonlinear
throughthicknessresidualstressdistribution.

A general method for predicting the manufacturing residual stresses and plastic

strains in coldformed steel members is proposed here. The procedure is founded on


common industry manufacturing practices and basic physical assumptions. The
primarymotivationforthedevelopmentofthismethodistodefinetheinitialstateofa
coldformedsteelmemberforuseinasubsequentnonlinearfiniteelementanalysis.The
derivation of the prediction method is provided for each manufacturing step, and the
predictions are evaluated with measured residual strains from existing experiments.
Theendresultofthemethodisintendedtobeaccessibletoawideaudienceincluding
manufacturers, designengineers, and the academic community. This method also has
the potential to compliment and improve Chapter A7.1.2 of the existing Specification
(AISIS100 2007), which currently allows for an increase in member strength from the

225

coldworkofformingeffectatcrosssectioncorners,butdoesnotdirectlyaccountforthe
influenceofthenonlinearthroughthicknesscornerresidualstressesortheinfluenceof
plastic strains and residual stresses from coiling, uncoiling, and flattening of the sheet
steel.

6.1 Stressstraincoordinatesystemandnotation
38B

Thestressstraincoordinatesystemandgeometricnotationusedintheforthcoming

derivationsaredefinedinFigure6.3.Thexaxisisreferredtoasthetransversedirection
1462H

and the zaxis as the longitudinal direction of a structural member. Crosssection


elementsarereferredtoaseithercornersorflats.Thesignconventionforstressand
strainispositivefortensionandnegativeforcompression.

roller dies

rx

sheet steel coil

Forming direction

Elevation View

t
y
rz

z
y
x
z

Section A-A

Figure6.3Stressstraincoordinatesystemasrelatedtothecoilingandcoldformingprocesses.

226

6.2 Predictionmethodassumptions
39B

Thefollowingassumptionsareemployedtodevelopthispredictionmethod:
a. Plane sections remain plane before and after coldforming of the sheet steel. This
assumptionpermitstheuseofbeammechanicstoderivepredictionequations.

b. The sheet thickness t remains constant before and after coldforming of the sheet
steel. A constant sheet thickness is expected after coldbending if the bending is
performed without applied tension (Hill 1950). Crosssection measurements
demonstratemodestsheetthinningatthecorners,wheretinthecornersistypicallyfive
percent less than in the flange and web (Dat 1980). This thinning is ignored here to
simplify the derivations, although a reduced thickness based on the plastic strain
calculationsinSection6.4couldbeusedifahigherlevelofaccuracyisrequired.
1463H

c. Thesheetneutralaxisremainsconstantbeforeandaftercrosssectioncoldforming.
Theoreticalmodelsusedinmetalformingtheorydopredictasmallshiftinthethrough
thicknessneutralaxistowardstheinsideofthecornerasthesheetplastifies(Hill1950).
This shift is calculated as six percent of the sheet thickness, t, when assuming a
centerline corner radius, rz, of 2.5t. A neutral axis shift of similar magnitude has been
observedinthenonlinearfiniteelementmodelresultsforthinpressbrakedsteelsheets
(Quachetal.2006).Thissmallshiftisignoredheretosimplifythederivations.

227

d.Thesteelstressstraincurveisassumedaselasticperfectlyplastic whencalculating
residual stresses. More detailed stressstrain models that include hardening are
obviously possible, but a basic model is chosen to simplify the derivations. The
implication of this assumption is that the residual stresses may be underestimated,
especially in corner regions where the sheet has yielded completely through the
thickness.

e. Plane strain behavior is assumed to exist during coiling, uncoiling, and flattening
(x=0) andduringcrosssectioncoldforming (z=0).

f. The steel sheet is fed from the top of the coil into the rollforming bed as shown in
Figure6.4a.Thisassumptionisconsistentwithmeasuredbendingresidualstressdata
146H

(see Section 6.6) and manufacturing setups suggested by rollforming equipment


1465H

suppliers(Figure6.1).TheauthordidobservethealternativesetupinFigure6.4b(sheet
146H

1467H

steel unrolling from the bottom of the coil) at a rollforming plant, suggesting that the
directionofuncoilingisasourceofvariabilityinmeasuredresidualstressdata.

228

Roll-forming bed

(a)

Sheet has residual


CONCAVE curvature
coming off the coil

(b)

Sheet has residual


CONVEX curvature
coming off the coil

Figure6.4Rollformingsetupwithsheetcoilfedfromthe(a)topofthecoiland(b)bottomofcoil.The
orientationofthecoilwithreferencetotherollformingbedinfluencesthedirectionofthecoilingresidual
stresses.

g. Membrane residual stresses are zero. Membrane residual stresses have been
measuredbyseveralresearchers(Ingvarsson1975;Dat1980;WengandPekz1990;De
BatistaandRodrigues1992;Kwon1992;Bernard1993;KeyandHancock1993),although
themagnitudesaresmallrelativetobendingresidualstresses(seeTable6.1).Membrane
1468H

residual stresses are experimentally determined by averaging the measured surface


strains on the two faces of a thin steel sheet. Given the variability inherent in these
measurementsitisdifficulttoknowiftheresultingmembranestresses(strains)arereal
orsimplyunavoidableexperimentalerror.

229

6.3 Derivation of the residual stress prediction


method
40B

The prediction method proposed here assumes that two manufacturing processes

contributetothethroughthicknessresidualstressesincoldformedsteelmembers:(1)
sheet coiling, uncoiling, and flattening, and (2) crosssection rollforming. Algebraic
equations for predicting the throughthickness residual stress and effective plastic
strains in corners and flats are derived here and then summarized in flowcharts in
Figure6.13andFigure6.17.
1469H

1470H

3.1 Residual stresses from sheet coiling, uncoiling, and flattening


8B

Coilingthesheetsteelafterannealingandgalvanizing,butpriortoshipment,may

yield the steel if the virgin yield strain, yield, is exceeded. If plastic deformation does
occur, a residual curvature will exist in the sheet as it is uncoiled. This residual
curvatureislockedintoastructuralmemberresultinginlongitudinalresidualstresses
as the sheet is flattened by the rollformers. This process of coiling, uncoiling with
residualcurvature,andflatteningisdescribedinFigure6.5.
147H

230

DETAIL A

Uncoiled with residual


curvature

Coiling

Flattened as sheet enters


the roll-formers

Change in curvature locks in


bending residual stresses in
final member

DETAIL A

Figure6.5Coilingofthesteelsheetmayresultinresidualcurvaturewhichresultsinbendingresidual
stressesasthesheetisflattened.

3.1.1 Coiling
160B

Thethroughthicknessstraininducedfromcoilingisrelatedtotheradiallocationof

thesheetinthecoilrx,withthewellknownrelationshipfrombeammechanics:

z
y

1
.
rx

(6.1)

zistheengineeringstrainthroughthethicknessyinthecoiling(longitudinal)direction
z.yvariesfromt/2tot/2,wheretisthesheetthickness.Theradiusassociatedwiththe
elasticplasticthresholdinitiatingthroughthicknessyieldingfromcoiling,rep,isderived
bysubstitutingz=yieldandy=t/2(outerfiberstrain)intoEq.(6.1):
1472H

rep =

t
2 yield

(6.2)

When the coil radius rx is greater than rep the sheet steel experiences only elastic
deformation on the coil. For sheetsteel rolled to a coil radius rx less than rep, through
thicknessyieldingwilloccurasshowninFigure6.6.
1473H

231

+yield

-yield

Figure6.6Longitudinalresidualstressdistributionfromcoiling.

Whenrx<rep,thedepthoftheelasticcorecisdefinedas:

c = 2rx yield t .

(6.3)

As the yielded sheet is uncoiled in preparation for the rollforming line, the sheet

3.1.2 Uncoiling
16B

steelspringsbackelasticallyresultinginachangeinthethroughthicknessstress.This
stressdistributionisdeterminedbyfirstcalculatingtheplasticcoilingmoment

coil
x

t 2 1

2
= yield (rx yield ) ,
2 3

(6.4)

and then applying an opposing moment elastically to simulate the removal of the
imposedradialdisplacement

uncoil
z

12 M xcoil y
=
.
t3

Afterthesheethasbeenunrolled,apermanentradiusofcurvaturewillstillexistifrx

(6.5)

3.1.3 Flattening
162B

waslessthanreponthecoil.Thispermanentradiusis

232

uncoil

rx

1
1 M xcoil

rx
EI

(6.6)

Steelsheetwithpermanentcurvaturefromcoilingispressedflatasthesheetentersthe
rollforming line. The longitudinal stresses resulting from flattening the sheet are
simply

zflatten = E

y
uncoil
x

(6.7)

3.1.4 Residualstressdistribution
163B

The total throughthickness longitudinal residual stress distribution due to coiling,

uncoiling,andflatteningispresentedinFigure6.7.
147H

+yield

uncoil
z

flatten
z

=
z

-yield
Coil

Uncoil

Flatten

Residual Stress

Figure6.7Predictedlongitudinalresidualstressdistributionfromcoiling,uncoiling,andflatteningofa
steelsheet.

Theresultingresidualstress,z, isselfequilibratingforaxialforcethroughthethickness
butcausesaresiduallongitudinalmoment.Section 6.6comparesthestressescausedby
1475H

thismomentwithsurfacestrains(stresses)measuredinexperiments.

The longitudinal residual stresses also will create transverse stresses across the

width of the coil, assuming plane strain conditions for an infinitely wide sheet.

233

Supportingtheplanestrainassumptionistheobservationthatwhiletheactualwidthof
the sheet is finite, it remains several orders of magnitude greater than the sheet
thickness. Under this assumption, and further assuming only elastic stresses, the
transversestressesare:

x = ( coil + uncoil +
z

flatten
z

) .

(6.8)

Poissons ratio, , is assumed here as 0.30 for steel deformed elastically. The through
thickness deformation from the uncoiling and flattening components will occur
elastically, and the coiling component will be at least partially elastic through the
thicknessfortherangeofsheetthicknessescommoninindustry.

3.2 Residual stresses from cross-section roll-forming


89B

A set of algebraic equations is derived here to predict the transverse and

longitudinal residual stresses created by rollforming a crosssection. Rollforming


residualstressesarecumulativewiththecoilingresidualstressesderivedinSection6.3.1
1476H

andprovideacompletepredictionoftheinitialstressstateofthemembercrosssection.
Therollformingresidualstressesareassumedtoexistonlyatthelocationoftheformed
corners, between the roller die reactions, as shown in Figure 6.8. Some yielding is
147H

expected to occur outside of the roller reactions as the stress distribution transitions
fromfullyplastictofullyelastic;however,thistransitionareaisnotconsideredhereto
simplifythederivation.

234

Assume bend is fully plastic


between roller dies

Roller Die (Typ.)

Figure6.8Coldformingofasteelsheet.

The engineering strain in the steel sheet, x, and the bend radius, rz, are related for

bothsmallandlargedeformationswiththestraincurvaturerelationship

1 x
=
rz
y

(6.9)

Thisgeometricrelationshipisvalidforelasticandplasticbendingofthesteelsheet.For
the small bend radii common in the coldformed steel industry (rz =2t to 8t), the steel
sheetyieldsthroughitsthicknessduringthecoldformingprocess.Thesteelsheetwill
reachthefullyplasticstressstateshownin Figure6.9asthecornerapproachesitsfinal
1478H

manufacturedradius.
y

-yield

+yield

Figure6.9Fullyplastictransversestressstatefromcoldforming.

235

After the sheet becomes fully plastic through its thickness, the engineering strain

continuestoincreaseastheradiusdecreases.Whenthefinalbendradiusisreachedand
theimposedradialdisplacementisremoved,anelasticspringbackoccursthatelastically
unloadsthecorner(seeFigure6.2).Thechangeinstressthroughthethicknessfromthis
1479H

elasticreboundisderivedwiththeplasticmomentforcecoupleshowninFigure6.10.
1480H

+yield
Fp
t/2

Fp

-yield

Figure6.10Forcecouple(Fpt)appliedtosimulatetheelasticspringbackofthesteelsheetafterthe
imposedradialdeformationisremoved.

Theplasticmomentiscalculatedwiththeequation

bend
z

2
t yield 1 t t yield t
= FP =
=
2
2
2
4

(6.10)

whichisthenappliedelasticallythroughthethicknesstosimulatethestressdistribution
fromelasticreboundofthesheetsteel:

xrebound

yield t 2

y
3 y
M zbend y 4
=
=
= yield .
1
t
I
1 t 3
12

The final transverse stress state is the summation of the fully plastic stress

(6.11)

distributionthroughthethicknessandtheunloadingstressfromtheelasticspringback
ofthecornerasshownin Figure6.11,wherexisthetransverseresidualstressthrough
148H

236

thethicknessfromthecoldformingofthecorner.Thisstressisnonlinearthroughthe
thickness and is selfequilibrating, meaning that axial and bending sectional forces are
absentinthexdirectionafterforming.
y

-yield

+1.5yield

+0.5yield

Plastic Bending

xbend

+yield

-yield
+yield

-0.5yield

-1.5yield
Elastic Springback

Transverse Residual Stress

xrebound

Figure6.11Coldformingofasteelsheetoccursasplasticbendingandelasticspringback,resultinginaself
equilibratingtransverseresidualstress.

Thetransverseresidualstresseswillcreatestressinthelongitudinaldirectionduetothe
assumedplanestrainconditions(seeSection6.2):
1482H

z = x .

(6.12)

ThePoissonsratio,,isassumedas0.30forsteeldeformedelasticallyand0.50forfully
plastic deformation. The longitudinal residual stresses through the thickness, z, are
determinedbasedontheseassumptionsasshownin Figure6.12.Longitudinalresidual
1483H

stress,z, isselfequilibratingforaxialforcethroughthethicknessbutcausesaresidual
longitudinal moment. This moment is hypothesized to contribute to the observed
longitudinal residual strains measured in experiments (refer to Section 6.6 for a
148H

comparisonofthispredictiontoactualmeasurements).

237

-yield

0.50

0.30

+yield
Plastic Bending

plastic xbend

-0.05yield

+1.5yield

-1.5yield

-0.50yield

+0.50yield
z

+0.05yield

Elastic Springback

Longitudinal Residual Stress

elastic xrebound

Figure6.12Plasticbendingandelasticspringbackfromcoldforminginthetransversedirectionresultin
longitudinalresidualstressesbecauseoftheplanestrainconditions.

A flowchart summarizing the proposed prediction method for residual stresses in

rollformed members is provided in Figure 6.13. Figure 6.13 explicitly demonstrates


1485H

1486H

how coiling, uncoiling, flattening, and rollforming contribute to the residual stresses
lockedintothecrosssectionduringmanufacturing.

238

Start

Flat

Corner
Flat or Corner?

End

No

t
2 yield
t
rx >
2 yield
rx

Yielding on the Coil?

No residual
stresses!

Yes, yields on coil


Yielding on the Coil?
No, remains elastic

c
t
y
2
2
c
c
< y<
2
2
t
c
y
2
2

+ yield

Yes

y
E
rx

zcoil =
Sheet Coiling

yield

c = 2rx yield
Sheet Coiling

M xcoil y
I

zuncoil =

t 2 1
2
1
M xcoil = yield (rx yield ) I = 1 t 3
12

2 3

xuncoil = elastic zuncoil

zflatten = E
uncoil

End

Yes

elastic = 0.30

xcoil = elastic zcoil

Sheet Uncoiling

rx

Sheet Flattening

y
rxuncoil

1
1 M xcoil

rx
EI

Sheet Flattening

elastic = 0.30

yield

0 y

t
2

+ yield t y 0

(For rz<8t)

zbend = plastic xbend

xrebound =

Sheet Uncoiling

elastic = 0.30

xflatten = elastic zflatten

bend
x

No

3 yield y
t

zrebound = elastic xrebound

Corner Bending

plastic = 0.50

t
t
y
2
2

Corner Rebound

End

elastic = 0.30

Figure6.13Flowchartsummarizingthepredictionmethodforresidualstressesinrollformedmembers.

239

6.4 Derivation of effective plastic strain prediction


method
41B

In the method proposed here, plastic strains occur from sheet coiling and cold

forming, and together with residual stresses describe the initial material state of the
member.Thegeneralstateofplasticstrainatapointcanbequantifiedbyusingthevon
Misesyieldcriterionextendedtoplasticdeformations(ChenandHan1988):

p =

2
3

2
1

+ 2 + 3
2

(6.13)

wherepistheeffectiveplasticstrain,and1, 2,and 3aretheprincipalstrains.Allofthe


strainsaretruestrains,whichmaybecalculatedfromtheengineeringstrainsvia:

1 = ln(1 + x ) , 2 = ln(1 + y ) , 3 = ln(1 + z ) ,

(6.14)

where x, y, zareintheCartesiancoordinatesystem(Figure6.3)andx,y,ziscoincident
1487H

withtheprincipaldirections.Truestrainsareemployedinsteadofengineeringstrainsto
accommodate the large deformations from plastic bending. Also, from a practical
standpoint,nonlinearFEcodessuchasABAQUS(ABAQUS2007a)requiretheengineer
toprovidetruestress,truestraininformation(aslargedeformationtheoryisemployed).
The steel sheet is assumed to remain incompressible while experiencing plastic
deformations,thereforewhencalculatingp

1 + 2 + 3 = 0 .

240

(6.15)

4.1 Effective plastic strain from sheet coiling


90B

Engineeringplasticstrains,asshowninFigure6.14,accumulateduringthecoilingof
148H

sheetsteelifthecoilingradiusrxislessthantheelasticplasticthresholdrep.
y

zp

Figure6.14Plasticstraindistributionfromsheetcoilingwitharadiuslessthanelasticplasticthresholdrep.

Theengineeringplasticstraindistributionfromcoilingis:

zp =

zp =

y
c
yield , y
rx
2

y
c
yield , y
rx
2

(6.16)

zp = 0 otherwise ,

wheretheelasticcore,c,isdefinedinEq.(6.3).Planestrainconditionsresultin1=0,and
1489H

2=3 via the incompressibility assumption of Eq. (6.15). Further, the Cartesian
1490H

coordinate system is coincident with the principal axes, resulting in the following true
principalplasticstrains:

1 = 0 , 2 = ln(1 + zp ) , 3 = ln(1 + zp ) .

(6.17)

Substituting the principal strains into Eq. (6.13) and simplifying leads to the through
149H

thicknesseffectiveplasticstrainfromcoiling

241

pcoiling =

2
ln (1 + zp ).
3

(6.18)

This plastic strain distribution, depicted in Figure 6.15, will exist at all locations in the
1492H

crosssection(cornersandflats)whenrxislessthantheelasticplasticthresholdrep.
pcoiling
y

Figure6.15Effectiveplasticstraininacoldformedsteelmemberfromsheetcoilingwhentheradiusrxis
lessthantheelasticplasticthresholdrep.

Theplasticstrainfromcoiling,pcoiling,willgenerallybemuchsmallerinmagnitudethan
theplasticstrainfromcrosssectioncoldforming,pbend,asdiscussedinfollowingsection.

4.2 Effective plastic strain from cross-section cold-forming


91B

Largetransverseplasticstrainsoccurthroughthethicknessofathinsteelsheetwhen

the sheet is permanently bent. The engineering plastic strain distribution from cold
formingisdescribedvia

xp =

y
rz

(6.19)

which assumes that the elastic core at the center of the sheet is infinitesimally small.
Thisassumptionisconsistentwiththesmallbendradiicommoninindustry(see 6.3.2).
1493H

242

PlanestrainconditionsandEq. (6.15)resultin3=0,2=1.Physicallytheseconditions
149H

implythatthesheetwillexperiencesomethinningatthelocationofcoldforming(see
Section6.2),butthetendencytoplasticallyshortenlongitudinallywillberesistedbythe
1495H

adjacent undeformed portion of the crosssection. As before, the Cartesian coordinate


systemiscoincidentwiththeprincipalaxes,resultinginthefollowingplasticprincipal
strains:

1 = ln(1 + xp ) , 2 = ln(1 + xp ) , 3 = 0 .

(6.20)

Substituting for the principal strains and simplifying, the effective plastic strain at a
coldformedcorneris:

bend
=
p

2
ln (1 + xp )
3

(6.21)

Thiseffectiveplasticstraindistributionisshownin Figure 6.16.Thedistributionexists


1496H

only at the coldbent locations in a crosssection and should be added to the coiling
plasticstraindistributioninFigure6.15.
1497H

Figure6.16EffectivevonMisestrueplasticstrainatthelocationofcoldformingofasteelsheet.

A flowchart summarizing the prediction method for effective plastic strains in roll
formedmembersisprovidedinFigure6.17.
1498H

243

Start

Flat

Corner
Flat or Corner?

End

No

t
2 yield
t
rx >
2 yield
rx

Yielding on the Coil?

No equivalent
plastic strains!

Yes, yields on coil


Yielding on the Coil?

No

No, remains elastic

Yes

End

y
yield
rx

Sheet Coiling

zp =

c
2
c
y
2
y

y
yield
rx

Sheet Coiling

c = 2rx yield

otherwise

bend
=
p

Yes

2
ln (1 + zp )
3

pcoiling =

2
ln (1 + xp )
3

xp =

t
t
y
2
2

Corner Bending

y
rz
End

Figure6.17Flowchartsummarizingthepredictionmethodforeffectiveplasticstrainsinrollformed
members

6.5 Employingthepredictionmethodinpractice:
quantifyingthecoilradiusinfluence
42B

The residual stress and plastic strain distributions derived for crosssection cold

forming(Sections6.3.2and6.4.2)arestraightforwardtocalculateiftheyieldstress,yield,
149H

150H

andthickness,t,ofthesheetsteelareknown.Thecoilingresidualstressesandplastic
strains are more difficult to calculate because the coil radius coinciding with the as
formed member, i.e., the radial location of the sheet, rx, is almost always unknown in

244

practice.However,rxcanbederivedinanaveragesensethough,sincetherangeofinner
and outer coil radii are known and the probability that a structural member will be
manufacturedfromacertainrxcanbequantified.

The relationship between coil radius, rx, and corresponding linear location S of the

sheetwithinthecoilcanbedescribedusingArchimedesspiral(CRC2003)

S=

(r
t

2
rinner .

(6.22)

Thespiralmaintainsaconstantpitchwithvaryingradii,wherethepitchisthethickness
ofthesteelsheet,t,asshownin Figure6.18,Listhetotallengthofsheetinthecoil,and
150H

rinner and router are the inside and outside coil radius, respectively. Asshipped outer coil
radii range from 24 in. to 36 in. and inner coil radii range from 10 in. to 12 in. These
rangesweredeterminedbytheauthorduringavisittoalocalrollformingplant.
S

rx
t

Start

S =0

End
S=L

Figure6.18Coilcoordinatesystemandnotation.

Archimedes spiral is used to describe the probability that the steel sheet will come

fromacertainrangeofradiallocationsinthecoil.Thecumulativedistributionfunction

245

(CDF),FR(rx)=probabilitythattheradiusislessthanrx,isobtainedbynormalizingSby
L

2
rx2 rinner
S
= FR (rx ) = 2
.
2
L
router rinner

(6.23)

Theprobabilitydensityfunction(PDF)ofrxiscalculatedbytakingthederivativeofFR(rx)

f R (rx ) =

2rx
dFR (rx )
= 2
.
2
drx
router rinner

(6.24)

Themeanvalueoftheradiallocationforagiveninnerandoutercoilradiiis

rx =

router
rinner

r r
2
f R (rx )rx drx = rinner + router inner outer
3
rinner + router

(6.25)

Thevarianceoftheradiallocationis

s R2 =

router

rinner

f R ( rx )(rx rx ) 2 dr =

(router rinner ) .
1 2
2
router + 4router rinner + rinner
18
(router + rinner )2

(6.26)

Thesestatisticsforrxcanthenbeusedwiththepredictionmethodforcoiling,uncoiling,
andflatteningresidualstressesandplasticstrainsdescribedinSections6.3.1and6.4.1.
1502H

1503H

Figure6.19summarizestheinfluenceofsheetthicknessandvirginyieldstressonthe
1504H

longitudinalresidualstressdistributionsinflatsandcorners.(Themethodproposedin
this chapter provides residual stresses and strains for the entire member, only the
longitudinal residual stresses are shown in Figure 6.19.) The solid lines in Figure 6.19
150H

1506H

arecalculatedusingthemeanvalue, rx =18.7in,fromEq.(6.25)assumingrinner=12in.and
1507H

router=24 in. The distributions with the dashed lines are calculated with rx sR, where
sR=3.4 in. is calculated with Eq. (6.26). The residual stresses are nonlinear through the
1508H

thicknessandhavedifferentshapesforflatsandcorners.Thestressmagnitudesatthe

246

outer fibers increase for thicker sheets and lower yield stresses. The accuracy of the
linearbendingresidual stressmodelcommonlyemployedinfiniteelementanalysesis
perhaps sufficient when yield stress is low and thickness is high (relatively), but for
typicalthicknesses(0.0346in.to0.0713in.)andyieldstress(50ksi)theassumptionofa
linear longitudinal stress distribution is not consistent with the mechanicsbased
predictionsinFigure6.19.
1509H

Longitudinal Residual Stresses


yield=50 ksi

yield=30 ksi
flat

t/2

corner

flat

yield=80 ksi

corner

flat

corner

t=0.0346 in.

-t/2

yield

yield

t/2

t=0.0713 in.

-t/2

yield

yield

t/2

t=0.1017
-t/2

yield

yield

Positive stress is tension, negative stress is compression


Figure6.19Influenceofsheetthicknessandyieldstressonthroughthicknesslongitudinalresidualstresses
(zdirection,solidlinesarepredictionsformeancoilradius,dashedlinesformeanradius+/onestandard
deviation).

247

6.6 Comparisonofpredictionmethodtomeasured
residualstresses
43B

The flat and corner residual surface strain measurements from 18 rollformed

specimens are used to evaluate the proposed residual stress prediction method. The
prediction methodprovidesthecompletethroughthicknesslongitudinalstrain(stress)
distributioniftheradiallocationinthecoilfromwhichthespecimenoriginatedinthe
coil, rx, is known. Since the radial coil location of the 18 specimens is unknown, rx is
statisticallyestimatedforeachspecimenusingthecoilradiusthatbestfitsthepredicted
surface strains to the measured surface strains from a specimen crosssection (for both
corners and flats). Once the best fit radial locations have been calculated, they are
examined to determine if their magnitude is rational when compared to typical inner
andouterdimensionsofasheetcoil.Althoughthiscomparisononlyprovidesapartial
evaluation of the prediction method, it is as far as one can go with the available data.
Qualitatively the prediction method is consistent with the more detailed through
thicknessfindings(KeyandHancock1993;Quachetal.2006).

6.1 Measurement statistics


92B

The mean and standard deviation of the residual stresses for the 18 rollformed

specimens used in this comparison are provided in Table 6.1. Positive membrane
150H

stresses are tensile stresses and positive bending stresses cause tension at y=t/2 (see
Figure 6.3 for coordinate system). The statistics demonstrate that both membrane and
15H

bending residual stress measurements are highly variable and that the membrane

248

stresses are small relative to the steel yield stresses. Details on the residual stress
measurements for each of the 18 specimens are described in a previous research
progressreport(MoenandSchafer2007b).
Table6.1Statisticsoftheresidualstressesinrollformedmembers

Element
Corners
Flats

Residual stress as %yield


Membrane
Bending
Mean
STDEV
Mean
STDEV
5.7
10.1
32.0
23.8
1.8
10.7
25.2
20.7

No. of
Samples
23
120

6.2 Mean-squared error (MSE) estimate of radial location


93B

Toexplorethevalidityofthepredictionmethod,theflatandcornerresidualstress

measurements from the 18 specimens are used to estimate the radial location rx from
which each specimen originated. These estimated radial locations are then used to
calculate the difference between the predicted and measured longitudinal residual
stresses.

6.2.1 MSEminimization
164B

The location of the specimen in the coil, rx, is estimated by minimizing the sum of the
meansquared errors (MSE) for the p=1,2,nq measurements taken around the cross
sectionoftheq=1,2,,18specimens
2

measured
predicted
pq

pq

.
rx ,q = arg min

yield , pq
p =1

nq

Bothcornerandflatmeasurementsareincludedintheminimization.

249

(6.27)

6.2.2 Bendingcomponentoflongitudinalresidualstressdistribution
165B

Thebendingcomponentofthepredictedresidualstressdistributionmustbeisolated

tocompare withthemeasuredvalues.Thetotalpredictedlongitudinal residualstress


distribution in the flats and corners of each crosssection is integrated to calculate the
sectionalmomentthroughthethickness

t
2
t

M x = z ydy .

(6.28)

Mxisthenconvertedintoapredictedouterfiberbendingresidualstresswhichcanthen
bedirectlycomparedtothemeasurements

pqpredicted

t
M x
2
=
.
I

(6.29)

6.2.3 EstimatedcoilradiiusingMSE
16B

Figure 6.20 demonstrates the meansquared error results for de M. Batista and
152H

Rodrigues Specimen CP1 (De Batista and Rodrigues 1992). The radial location that
minimizesthepredictionerroris1.60rinnerinthiscase,andissummarizedinTable6.2for
153H

all18rollformedspecimensconsidered.Theestimatedradiallocationsfallwithinthe
rangeofinnerandoutercoilradiiassumedintheprediction(rinner to2.40rinner)exceptfor
Dat RFC13 which is slightly outside the range at 2.45rinner. The MSE radial location
cannotbedeterminedinthethreeBernardspecimens(Bernard1993)sincethebending
residual stresses in the flats are predicted to be zero. These three specimens are cold

250

formedsteeldeckingwithathinsheetthicknesstrangingfrom0.022in.0.0400in.anda
relativelyhighyieldstressyieldrangingfrom87ksito94ksi.Inthiscase,thecoilingand
uncoiling of the steel sheet will occur elastically as demonstrated in Figure 6.19.
154H

MeasuredbendingresidualstressmagnitudesintheflatsoftheBernardspecimensare
onaverage0.03yieldwhichisconsistentwiththepredictionmethod.

8
7
6

MSE

5
4
3
2
1
0

1.5

2.5
rx /rinner

3.5

Figure6.20ThemeansquarederrorofthepredictedandmeasuredbendingresidualstressesfordeM.
BatistaandRodrigues(DeBatistaandRodrigues1992),SpecimenCP1isminimizedwhenrx=1.60rinner.

251

Table6.2Radiallocationinthecoilthatminimizesthesumofthemeansquarepredictionerrorforroll
formedmembers
Researcher

rx estimate

Specimen

de M. Batista and Rodrigues (1992)


de M. Batista and Rodrigues (1992)
Weng and Pekz (1990)
Weng and Pekz (1990)
Weng and Pekz (1990)
Weng and Pekz (1990)
Weng and Pekz (1990)
Weng and Pekz (1990)
Weng and Pekz (1990)
Weng and Pekz (1990)
Dat (1980)
Dat (1980)
Bernard (1993)
Bernard (1993)
Bernard (1993)
Abdel-Rahman and Siva (1997)
Abdel-Rahman and Siva (1997)
Abdel-Rahman and Siva (1997)
rinner=10 in., router=24 in.

in.
12.0
16.0
18.0
11.0
14.5
13.0
19.5
15.0
23.0
16.0
20.0
24.5
N/A
N/A
N/A
16
16
13

CP2
CP1
RFC13
RFC14
R13
R14
P3300
P4100
DC-12
DC-14
RFC14
RFC13
Bondek 1
Bondek 2
Condeck HP
Type A - Spec 1
Type A - Spec 2
Type B - Spec 1

N/A coiling residual stresses are predicted as zero

6.3 Statistical variations between measurements and predictions


94B

The predicted radial locations in Table 6.2 are now used to calculate the statistical
15H

variations between the experiments and predictions. The bending residual stresses in
the 18 rollformed members are calculated using the MSEpredicted radial location rx
with the residual stress prediction method summarized in Figure 6.13. The bending
156H

component of the residual stress prediction is then obtained with Eq. (6.29). The
157H

difference between the predicted and measured residual bending stresses, epq, for the
p=1,2,,nmeasurementstakenaroundthecrosssectionoftheq=1,2,,18specimensis
calculatedas

measured
predicted
pq
pq
e pq =
yield , pq

252

(6.30)

The error histogram for the flat crosssectional elements in Figure 6.21a demonstrates
158H

that the mean difference e is near zero with a standard deviation se=0.15yield. The
scattergram in Figure 6.21b demonstrates the strength of the correlation between the
159H

measurements and predictions in the flats; the solid regression line passes nearly
through zero (yintercept=0.05yield) and has nearly a unit slope (m=0.92). Also, the
majorityofthedatalieswithinonestandarddeviationoftheestimate,denotedasthe
dashedlinesinthefigure.Itisconcludedthatthepredictionmethodisconsistentwith
themeasureddataintheflats.

The corner element error histogram in Figure 6.22a shows a negative bias of e=
1520H

0.16yield meaning that the predicted residual stresses are generally higher than the
measured values. The standard deviation of the error is large (se=0.19yield) but is less
than the standard deviation of the corner residual stress measurements in Table 6.1
152H

(sm=0.24yield). This demonstrates a greater match between the measurements and


predictions,althoughmorecornerresidualstressmeasurementsareneededtoimprove
the strength of this comparison. The scattergram in Figure 6.22b highlights the
152H

variability in the measured corner data, especially in the region corresponding to


predicted=0.4yield,wherebendingresidualstresses(strains)varyfrom0to0.7yield.

253

40

e = 0.03 yield

30

e = 0.15 yield

0.7
0.6
0.5
/ yield

25

measured

20
15

0.4
0.3
0.2

Observations

0.8

35

0.1

10

-0.1

0
-1

-0.5

0
(

measured

0.5

-0.2

0.1

0.2

predicted

)/ yield

0.3

0.4

predicted

(a)

0.5

0.6

0.7

0.8

/ yield

(b)

Figure6.21(a)Histogramand(b)scattergramofbendingresidualstresspredictionerror(flatcross
sectionalelements)for18rollformedspecimens.

40
0.8
0.7

e = 0.16 yield

25

e = 0.19 yield

0.6
0.5
/ yield

30

measured

20
15

0.4
0.3
0.2

Observations

35

0.1

10

-0.1

0
-1

-0.5

0
(

measured

0.5

-0.2

predicted

)/ yield

(a)

0.1

0.2

0.3

0.4

predicted

0.5

0.6

0.7

0.8

/ yield

(b)

Figure6.22(a)Histogramand(b)scattergramofbendingresidualstresspredictionerror(cornercross
sectionalelements)for18rollformedspecimens.

6.7 Discussion
4B

The residual stresses and strains predicted with this method (Section 6.3 for stress,
1523H

Section 6.4 for strain) form the initial material state in thecrosssection. In design, this
1524H

initialmaterialstateissometimesconsideredthroughthesocalledcoldworkofforming

254

effect, where the yield stress of the material is increased above the virgin yield stress,

yield, to account for the working of the corners. For onedimensional stressstrain this
concept is expressed as shown in Figure 6.23, where working the corners results in a
152H

residual plastic strain, p, such that when the section is reloaded the stress at which
yieldingreinitiates,ey,isgreaterthanthevirginyieldstress,yield.Ifnoresidualstresses
existed the apparent increase in the yield stress from yield to ey can be significant.
However, as Figure 6.19 illustrates, working the corners also contributes to residual
1526H

stresses,ers,andtheseresidualstressesmaydecreasetheapparentyieldstress.

Thepredictionmethodpresentedhereinprovidesamorenuancedunderstandingof

thecoldworkofformingeffects.Theresidualplasticstrainsmayincreasetheapparent
yield stress, but those strains vary through the thickness and have contributions from
both transverse and longitudinal strains. Further, residual stresses follow their own
relatively complicated distribution through the thickness. In a multiaxial stress state
using the von Mises yield criterion, Figure 6.23 is enforced for the effective stress
1527H

effectivestrainpairfor everypoint inthecrosssection.Asaresult,theapparentyield


stressuponloadingvariesthroughthethicknessandisinfluencedbyboththeresidual
stresses and strains. Even under simple loading conditions (e.g., compression) a cold
formed member undergoes plate bending well in advance of collapse, so the strains
demandedofthematerialalsovarythroughthethicknessandaroundthecrosssection.
Whileitisindeedpossibletomodelsucheffectsinafiniteelementanalysis,assuming
these effects can be collapsed into a generic increase in the yield stress for the entire
sectionasiscurrentlydoneindesignwouldseemtobeanoversimplification.

255


Virgin steel

ey

yield

Coldformed
steel

ers

Apparent
yield
stress

Effective residual
stress

yield + p

Figure6.23Definitionofapparentyieldstress,effectiveresidualstress,andeffectiveplasticstrainasrelated
toauniaxialtensilecoupontest.

Implementationoftheresidualstressesandinitialplasticstrainsintoacommercial

finite element program such as ABAQUS, where the member is modeled using shell
elements, is relatively straightforward. The number of throughthickness section
(integration) points must be increased to resolve the nonlinear throughthickness
residualstressandstraindistributions.Theresidualstressesandstrainspredictedherein
can be relatively large. Further, conventional loading (e.g., compression, majoraxis
bending)maycauseloadingorunloadingoftheseinitialstressesatagivenpointinthe
crosssection. As a result, the hardening rule: isotropic, kinematic, or mixed can have
practicaldifferencesintheobservedresponseevenwhentheappliedloadsthemselves
arenotreversing.

Forthissituation,kinematichardening,whichapproximatestheBauschingereffect,

providesamoreconservativemodeloftheanticipatedmaterialbehaviorthanisotropic
hardening. However, to model kinematic hardening the location of the center of the
yieldsurfaceinstressspace(alsoknownasthebackstress)mustbedeterminedforeach

256

point in the crosssection at the end of the manufacturing process. This location is a
functionoftheextentofyielding,intheexampleofFigure6.23,thebackstresswouldbe
1528H

the 1, 2, 3 triad that results in the effective stress increasing from yield to ey.
Unfortunately, the elasticperfectly plastic assumption used to predict residual stresses
herein does not directly allow for the calculation of the backstress. However, the
effective plastic strain may be used to approximate the backstress as provided in
AppendixG.Furtherexaminationofthepredictedresidualstressandstrainsandtheir
1529H

impact on the peak strength and collapse response of coldformed steel members in
nonlinearfiniteelementanalysisiscurrentlyunderway,includingtheworkpresentedin
Section7.2.
1530H

6.8 Acknowledgements
45B

The development of this residual stress prediction method would not have been

possible without accurate information about the manufacturing process of sheet steel
coilsandcoldformedsteelmembers.ThankstoClarkWesternBuildingSystems,Mittal
Steel USA, and the ColdFormed Steel Engineers Institute (CFSEI) for their important
contributions to this research, especially Bill Craig, Ken Curtis, Tom Lemler, Joe
Wellinghoff,EzioDefrancesco,JeanFraser,NarayanPottore,andDonAllen.

257

Chapter 7
Nonlinear finite element modeling of
cold-formed steel structural members
6B

Commercial finite element programs provide a means for realistic collapse


simulation of coldformed steel structural members. Thin shell finite element
formulationsprovidedinABAQUS(e.g.,theS9R5elementdiscussedin Chapter2)are
153H

designed to capture the sharp folds and throughthickness yielding characteristic of


coldformed steel beams and columns at their ultimate limit state. Robust solution
algorithms are available to predict unstable, geometrically nonlinear collapse. The
abilitytodefinetheinitialstateofamember,includinggeometricimperfectionsandthe
effectsofresidualstressesandinitialplasticstrainsfromthemanufacturingprocess,is
alsofeasible.Caremustbetakenthoughwithcomputationalresultssincetheyareoften
sensitivetomodelinginputsandassumptions.Itisprudenttostudythesesensitivities
andvalidateaspecificmodelingprotocolwithknownexperimentresultsbeforetrusting
theprotocoltoconsistentlyproducephysicallyrealisticresults.

258

This chapter begins with preliminary nonlinear finite element studies of stiffened
elements(i.e.,asimplysupportedplate,see Figure3.1fordefinition)withandwithout
1532H

holes, which are designed to gain experience with available ABAQUS nonlinear finite
elementsolutionmethods.Theinfluenceofimperfectionsonstiffenedelementsisalso
evaluated, and the throughthickness yielding patterns of a stiffened element (i.e.,
effectivewidth)withandwithoutaholearecompared.Theconclusionsreachedfrom
thispreliminaryworkareusedtoguidethedevelopmentandvalidationofanonlinear
finite element modeling protocol which is needed in Chapter 8 to explore the Direct
153H

StrengthMethodformemberswithholes.

7.1 PreliminarynonlinearFEstudies
46B

Exploratory nonlinear finite element studies are conducted in this section to gain

experience with ABAQUS input parameters and solution controls. All studies are
focused on the simulation of a stiffened element loaded unixaxially to collapse, and
specificattentionispaidtothemodelingofastiffenedelementwithahole.Experience
gainedfromsolvingthishighlynonlinearproblemwillbevaluablewhenimplementing
the larger simulation studies on full coldformed steel members with holes in Section
7.2.
1534H

1.1 Finite element modeling definitions


95B

The stiffened element is modeled with ABAQUS S9R5 thin shell finite elements,

where the plate dimensions are h=3.4 in. and L=27.2 in. (see Figure 3.2 for plate
153H

dimension definitions) and the plate thickness t is 0.0346 in. (These dimensions are

259

specifically chosen to be consistent with the flat web width and thickness of an SSMA
362S16233 structural stud.) Coldformed steel material properties are assumed as
E=29500 ksi and =0.30. Material nonlinearity is simulated in ABAQUS with classical
metal plasticity theory, including the assumption of a von Mises yield surface and
isotropic hardening behavior. The nonlinear plastic portion of the true stressstrain
curveshownin Figure7.1wasobtainedfromatensilecoupontest(Yu2005)andinput
1536H

intoABAQUStodefinethelimitsofthevonMisesyieldsurfaceasafunctionofplastic
strain.

70

60

true plastic stress, ksi

50

40

True stress True strain


ksi
33
0.0000
34.31
0.0016
42.14
0.0259
51.35
0.0700
60.86
0.1407
65.5
0.1988

30

20

10

0.02

0.04

0.06

0.08

0.1
0.12
true strain

0.14

0.16

0.18

0.2

Figure7.1Truestressstraincurvederivedfromatensilecoupontest(Yu2005)

Theboundaryconditionsofthestiffenedelementaresummarizedin Figure7.2.
1537H

Theplateissimplysupportedaroundtheperimeterwithsidesfreetowave.Thenodes
attheloadededgesoftheplatearecoupledtodisplacetogetherlongitudinally(inthe1

260

direction), which prevents local failure modes of the plate at the loaded edges. The
nodalcouplingisprovidedwithanequationconstraintinABAQUS.
Loaded edge coupled to move
together in 1 using equation
constraint (all u are equal)

Transverse midline
restrained in 3 (w=0)

Perimeter restrained in 2 (v=0),


unloaded edges free to wave
Longitudinal midline
restrained in 1 (u=0)

2
Loaded edge coupled to move
together in 1 using equation
constraint (all u are equal)

Figure7.2Simplysupportedboundaryconditionswithequationconstraintcouplingatloadededges

Two types of loading conditions, uniform load and uniform displacement, are
considered as shown in Figure 7.3. The uniform compressive load is applied as
1538H

consistent nodal loads on the plate edge. The magnitude of the uniform load is
representedbytheparameter,whichisanaccumulationofloadstepsautomatically
determined by ABAQUS. is large when the NewtonRaphson algorithm converges
quickly(alongthelinearbranchoftheloaddisplacementcurve)andadjuststosmaller
incrementsasequilibriumbecomesmoredifficulttoachieve(nearthepeakoftheload
displacementcurve).Fortheuniformdisplacementcase,thetotaldisplacementofthe
plate edges is applied over 100 steps, where the maximum displacement increment at
eachstepissetto=0.0145t.

261

P
h

h
(b)

(a)

Load scaling factor


P Unit force
h Loaded width of the plate
Displacement of plate edge

Figure7.3Applicationof(a)uniformloadand(b)uniformdisplacementtoastiffenedelement

Initial geometric imperfections are imposed based on the fundamental elastic


buckling mode of the stiffened element (see buckled shape in Figure 7.3). The
1539H

magnitudeoftheimperfectionsischosenbasedonaprobabilistictreatmentdeveloped
for coldformed steel members (Schafer and Pekz 1998). Since the stiffened element
considered here is chosen to be consistent with the web of a structural stud, a Type 1
(local buckling) imperfection is assumed as shown in Figure 7.4. The maximum
1540H

magnitudeoftheimperfectionfieldisselectedsuchthatthereisa50percentchancethat
a randomly occurring imperfection in the plate, , will have a magnitude less than d1,
i.e.,P(<d1)=0.50.Forthisprobabilityofoccurrence,theinitialimperfectionfieldofthe
stiffenedelementisscaledtod1/t=0.34.

262

Cross section with initial


geometric imperfections

d1

Figure7.4Type1imperfection(SchaferandPekz1998)

1.2 ABAQUS nonlinear solution methods


96B

Twononlinearsolutionmethods,themodifiedRiksmethodandaNewtonRaphson

techniquewithartificialdamping,areavailableinABAQUStosolvedifficultnonlinear
problems.ThemodifiedRiksMethod(i.e.,*STATIC,RIKSinABAQUS),wasdeveloped
in the early 1980s and enforces an arc length constraint on the NewtonRaphson
incremental solution to assist in the identification of the equilibrium path at highly
nonlinearpointsalongtheloaddeflectioncurve.Thismethodisdiscussedextensively
in several publications (Crisfield 1981; Powell and Simons 1981; Ramm 1981; Schafer
1997; ABAQUS 2007a). Another solution option is a NewtonRaphson technique (i.e.,
*STATIC,STABILIZEinABAQUS)whichaddsartificialmassproportionaldampingas
localinstabilitiesdevelop(thatis,whenchangesinnodaldisplacementsincreaserapidly
over a solution increment) to maintain equilibrium (Yu 2005; ABAQUS 2007a). Local
instabilitiesnearpeakloadarecommonincoldformedsteelmembers,suchaswhena
thinplatedevelopsatafoldlinepriortocollapse.

263

Inthisstudy,thestiffenedelementdescribedinSection7.1.1isloadedtocollapse
154H

in ABAQUS employing the modified Riks method with uniform loads applied
uniaxially (see Figure 7.3a) and then with the artificial damping solution method
1542H

employing uniform displacements (see Figure 7.3b). (Either method is capable of


1543H

solving problems with applied loads or applied displacements.) The goal of this
preliminary study is to gain experience with the solution controls for each method.
Additional background information pertaining to the ABAQUS implementation of the
artificial damping method is also discussed to provide specific guidance (and raise
futureresearchquestions)onitsproperuse.

1.2.1 ModifiedRikssolution
167B

The loaddisplacement curves and deformed shapes (at peak load) of the
stiffened element solved with the modified Riks method are provided in Figure 7.5.
154H

Different postpeak equilibrium paths were obtained by varying max, the maximum
loadincrementlimitfortheABAQUSautomaticstepselectionalgorithm.Theexistence
of multiple solutions is consistent with a plate containing periodic geometric
imperfections,sinceeachhalfwavelengthhasanequalchanceofdeforminglocallyinto
aplasticfailurezone.
Althoughtherewereseveraldifferentpostpeakpathsobserveddependingupon
the choice of max, the primary failure mechanism for the plate was a sharp yieldline
fold occurring transversely across the plate. Figure 7.6 demonstrates that this folding
154H

occursatthecrestofthebuckledhalfwaveoftheinitialgeometricimperfectionfield;in

264

this case the failure mechanism of the plate is linked to the initial imperfection shape.
The quantity and location of the plastic folds influenced the overall ductility of the
stiffenedelement(i.e.,theareaundertheloaddisplacementcurve).Asthenumberof
folds increase, the postpeak strength and ductility of the plate increase. The peak
compressiveloadofthestiffenedelementwasnotsensitivetochangesinmax.

Failure modes and


associated Riks step sizes

0.9
0.8

0.8
max = 0.70

0.7
0.6

P/Py

0.6

max = 0.10, 0.20, 0.25,


0.30, 0.40, 0.60

0.5

0.4

max = 0.50

0.4
0.3

0.2

0.2

P
h

max = 0.35

0.1
0

Initial imperfection shape


(scale exaggerated)

0.2

0
0.25

0.5

0.4

P
h

0.75

0.6
1

0.8
1.25

1.5

max = 0.05,0.15

1.75

/t

Figure7.5ModifiedRiksmethodloaddisplacementsolutionsandfailuremodes

ultimate limit state


fold line (typ.)

initial state with


assumed
imperfections

(imperfection and deformation magnitudes not to scale)

Figure7.6Correlationbetweeninitialimperfectionshapeandfoldlinelocationsatfailure

265

1.2.2 Artificialdampingsolution
168B

7.1.2.2.1 Backgroundonartificialdampingsolutionmethod
208B

Artificial mass proportional damping is employed in ABAQUS to alleviate local

instabilities in the *STATIC, STABILIZE solution method. The global equilibrium


equationsateachdisplacementstepcanbewrittenas:

P F D = 0

(7.1)

where P is the vector of applied external forces, F is the vector of calculated internal
forces, and D is the vector of viscous damping forces. The damping force vector D is
calculatedateachstepbasedonthefollowingrelationship:

D = (cM )v

(7.2)

where c is the damping ratio, M is an artificial mass matrix calculated with a unit
materialdensity,andvrepresentsthechangeinnodaldisplacementsdividedbythesize
of the time step selected by ABAQUS. v is called the nodal velocity in ABAQUS
since the dimensions are length/time, which makes v sensitive to the definition of
time.Inthisstudy,thetotaltimeisselectedasoneunitandthemaximumtime
stepallowedis0.01units.Ifthetotaltimeischosenas100unitsandthemaximum
time step as 1 unit, it seems that the magnitude of the damping forces D would
change.Followingthesameargument,themagnitudeofvisimpactedbythechoiceof
unitsfortheproblem(feet,inches,meters,mm)sincevhasdimensionsoflength/time
units.Futureworkisneededtoevaluatetheinfluenceoftimeandlengthunitsonthe
calculationofthenodalvelocityv.Theevaluationoftheartificialdampingsolution

266

sensitivitytothemagnitudeanddistributionofmassinamemberisalsoanotherfuture
researchtopic.
When the solution is stable, changes in nodal displacements are small and
viscousdampingisnegligible.Whenlargechangesindisplacementsoccurbetweentwo
consecutiveloadsteps(asinthecaseofalocalinstability),dampingforcesareappliedto
helpmakeupthedifferencebetweenPandF.vmayonlybehighforcertainlocations
inthemember,andthereforedampingwillonlybeappliedthere.ABAQUSprovides
both automatic and manual options for selection the damping factor c; if c is chosen
manually,ABAQUSrecommendsthatitshouldbechosenasasmallnumbersincelarge
damping forces can add too much artificial stiffness to the system, producing an
unreasonable solution. When the automatic option is selected, ABAQUS finds c such
thatthedissipatedenergytototalstrainenergyratioafterthefirstloadstepisequalto
2.0x104.

1.2.3 Artificialdampingresults
169B

TheartificialdampingsolutionsforthestiffenedelementarepresentedinFigure
1546H

7.7. Loaddeformation results pertaining to both manually and ABAQUSselected


damping factors are plotted, demonstrating that the magnitude of the damping
parameter c influences the postpeak response and causes the prediction of several
differentloadpaths(inasimilarwaytohowmaxaffectedthemodifiedRikssolutions).
Peakloadisnotsensitivetothequantityofdampinginthiscase,andisconsistentwith
themagnitudepredictedwiththemodifiedRiksmethod.

267

0.9
0.8

ABAQUS chooses this


path (c=0.0162)

0.8

Failure modes and


associated damping
factor c

0.7
0.6

P/Py

0.6

c=0.005

0.5

0.4

0.4
0.3

0.2

0.2

c=0.000, 0.0162

c=0.050

0.1
0

Initial imperfection shape


(scale exaggerated)

P
h

Displacement
0.2
Control

0.25

0.5

0.4

P
h

0.75

0.6
1

0.8
1.25

1.5

1
1.75

/t

Figure7.7Artificialdampingloaddisplacementsolutionsandfailuremodes

1.3 ABAQUS nonlinear solution controls


97B

Section 7.1.2 summarizes the preliminary experiences gained using ABAQUS


1547H

nonlinear solution methods to determine the ultimate strength of stiffened elements.


Equilibriumpathsandfailuremodescanbesensitivetosolutioncontrols,althoughthe
peakresistingloadoftheplatewasconsistentlypredicted.Thenonlinearsolutionofa
stiffened element with a slotted hole is attempted with the modified Riks method
(*STATIC, RIKS), the default NewtonRaphson solution algorithm (*STATIC), and
NewtonRaphson with artificial damping (*STATIC, STABILIZE) solution methods
availableinABAQUS.Thegoalofthestudyistodetermineasetofsolutionscontrols

268

(loadstepsize,dampingfactor,convergencelimits)thatiscapableofcapturingthepost
peakloaddisplacementresponseofastiffenedelementwithahole.

1.3.1 Problemdescription
170B

The stiffened element described in Section 7.1.1 is considered in this study. A


1548H

single slotted hole is placed at the midlength of the plate and centered between the
unloadededges.Theslottedholehasdimensionsofhhole=1.5in.,Lhole=4in.,andrhole=0.75in.
(seeFigure3.2forholedimensiondefinitions).
1549H

The boundary conditions of the stiffened element were initially assumed to be


simplysupported with both transverse and longitudinal plate midlines restrained and
theloadededgenodescoupledwithconstraintequationsasdescribedin Figure7.2.It
150H

wasoftenobservedthattheconstraintsusedtoenforceuniformdisplacementsin1(u)at
the loaded edges and the transverse midline restraint in 2 (w=0) were contributing to
solutionconvergenceproblems,andthereforeanalternativesetofboundaryconditions
wasdevelopedasshowninFigure7.8.
15H

Restraint at center of
loaded edge in 3 (w=0)
Loaded edge coupled to move
together in 1 using rigid body
formulation in ABAQUS
(all u are equal)
Perimeter restrained in 2 (v=0),
unloaded edges free to wave
Longitudinal midline
restrained in 1 (u=0)
Restraint at center of
loaded edge in 3 (w=0)

2
Loaded edge coupled to move
together in 1 using rigid body
formulation in ABAQUS
(all u are equal)

Figure7.8Stiffenedelementboundaryconditionswithrigidbodycouplingatloadededges

269


Acomparisonofthegeometricimperfectionsassumedforthestiffenedelement
withandwithouttheholeisprovidedin Figure7.9.d1/t=0.34isusedtoscaletheinitial
152H

imperfection field of the plate. This magnitude corresponds to a probability of


occurrenceofP(<d1)=0.50(seeSection 7.1.1fordetails).Thepeakloadofthestiffened
153H

element is sensitive to initial geometric imperfections, and therefore it is important to


considerthesameimperfectionshapewhencomparingtheloaddisplacementresponses
ofthestiffenedelementwithandwithoutahole.Theimperfectionshapeisimposedon
thestiffenedelementwiththeslottedholebymappingthebuckledmodeshapetonodal
coordinatesusingcustomMATLABcode(Mathworks2006).
fundamental
buckling mode of
plate
initial geometric
imperfection fields

fundamental buckling
mode mapped to plate
with slotted hole

Figure7.9Initialgeometricimperfectionfieldusedforthestiffenedelementwithandwithoutahole

EightexploratoryABAQUSmodelsareevaluatedinthisstudy,eachsolvingthe
samenonlinearproblemofastiffenedelementwithaslottedholecompresseduniaxially
until failure as shown in Figure 7.10. Each model employs a different combination of
154H

ABAQUSsolutioncontrolsandboundaryconditionsassummarizedinTable7.1.
15H

270

Figure7.10Deformationatultimateloadofastiffenedelementwithaholeloadedincompression.Thecommonfailuremechanismismaterialyieldingadjacent
totheholefollowedbyplatefolding.

Table7.1Summaryofnonlinearfiniteelementmodelsandassociatedsolutioncontrols
*STATIC
Model

ABAQUS Method

initial
step size

total
time

min step
size

RIKS1
RIKS2
RIKS3
STATIC1
STATIC2
STAB1
STAB2
STAB3

*STATIC, RIKS
*STATIC, RIKS
*STATIC, RIKS
*STATIC
*STATIC
*STATIC, STABILIZE
*STATIC, STABILIZE
*STATIC, STABILIZE

------0.01
0.01
0.01
0.01
0.01

------1
1
1
1
1

------1.00E-20
1.00E-20
1.00E-20
1.00E-20
1.00E-20

max
step
size
------0.01
0.01
0.01
0.01
0.01

*STATIC, RIKS
Solution Controls
initial
residual
damping
iteration limits
max step
step
limits
factor
size

size
Io
Ir
Ic
c
Rn
0.05
0.2
4 (D) 8 (D) 16 (D) 0.005 (D)
--(D)
(D)
4 (D) 8 (D) 16 (D) 0.005 (D)
--0.05
0.05
8
16
33
0.005 (D)
------8
16
33
0.005 (D)
------8
16
33
0.005 (D)
------8
16
33
0.005 (D)
0.0162
----8
16
33
0.1
0.0162
----8
16
33
0.005 (D)
0.0162

271

line
search
NO
NO
NO
NO
YES
NO
NO
NO

Loading Type and Boundary Conditions


uniform load, equation constraint
uniform load, equation constraint
imposed displacement, rigid body constraint
uniform load, equation constraint
uniform load, equation constraint
imposed displacement, rigid body constraint
imposed displacement, rigid body constraint
uniform load, rigid body constraint

1.3.2 ModifiedRiksmethodsolutioncontrols
17B

The RIKS1 and RIKS2 finite element models are loaded with a uniformly
distributed load at both ends as shown in Figure 7.3(a), where equation constraints
156H

couple the loaded edge nodes (see Figure 7.2). The initial and maximum load step
157H

magnitudesaredefinedforRIKS1basedonexperiencegainedfromthestudyinSection
7.1.2.1. The RIKS2 model allows ABAQUS to select all load stepping parameters
158H

automatically.

1
RIKS1
RIKS2

0.8
0.6

cannot move
past peak load

0.4

P/Py,g

0.2

compression

tension

-0.2

-0.4

-0.6

-0.8
-1
-2

Initial imperfection shape


(scale exaggerated)

P
b

-1.5

-1

-0.5

/t

0.5

1.5

P
b

Figure7.11LoaddisplacementcurvefortheRIKS1andRIKS2modelsshowingdirectionreversalalong
loadpath

The loaddisplacement responses from the RIKS1 and RIKS2 models are
compared in Figure 7.11. For both models, ABAQUS does not capture the peak load
159H

andrevertsbackalongtheequilibriumpathuntiltheplateisloadedtofailureintension!
TheABAQUSTheoryManualstatesthatthistypeofdirectionswitchispossiblewhen

272

the equilibrium path exhibits very high curvature (ABAQUS 2007a). The ABAQUS
messagefiles(.msg)for thesemodelsreportthatthemomentresidualsaretoohighat
theloadededgenodesandatnodesalongthetransversemidlineoftheplate,suggesting
that these boundary conditions are contributing to the convergence difficulties for the
solution.
TheRIKS3modelisloadedwithimposeddisplacementsatbothendsasshown
inFigure7.3(b),wherethemidlineconstraintisremovedandtheloadededgenodesare
1560H

coupledwitharigidbodyconstraintinsteadofanequationconstraintinABAQUS(see
Figure 7.8). According the ABAQUS Analysis Users Manual, only the reference node
156H

governing the motion of the rigid body is involved in element level calculations. This
improves computational efficiency and releases the solution algorithm from the force
andmomentresidualminimizationconstraintsforallnodesintherigidbodyexceptthe
referencenode.
The solution results from the stiffened elements loaded with consistent nodal
loads(RIKS1,RIKS2)andimposeddisplacements(RIKS3)arecomparedin Figure7.12.
1562H

Before yielding occurs, the three models produce nearly identical loaddisplacement
results. As yielding initiates, the RIKS3 model predicts a peak load and postpeak
response for the stiffened element. This comparison demonstrates that imposed
displacements and rigid body constraints (in contrast to applied loads and equation
constraints)improvethechancesforconvergenceinthiscase.

273

RIKS1, RIKS2
RIKS3

P/Py,g or P/Py,g

0.8
0.6

0.39
0.8

0.4

0.37

0.2

0.6
0.35
0.24

0
-0.2
-0.4

Riks Method finds post-peak path


when plate is compressed with
imposed displacements

0.29

0.34

0.4
Riks Method retreats back along
elastic load path when plate is
loaded with consistent nodal
loads

0.2

-0.6
-0.8

P
h

-1
-2

0
-1.5

0.2
-1

0.4
-0.5

0.6
0

/t

0.5

0.8
1

1.5

P
1h
2

Figure7.12RIKS1andRIKS2modelsexperienceconvergenceproblemsandreturnalongtheloadingpath,
theRIKS3modelsuccessfullypredictsapeakloadandfindsapostpeakloadpath

1.3.3 NewtonRaphsonmethod
172B

TheSTATIC1andSTATIC2modelsemploytheNewtonRaphsonalgorithmwith
uniform displacements at the loaded edges imposed with equation constraints (see
Figure 7.2). The stepping parameters are chosen to ensure at least 100 increments are
1563H

achieved before completion of the simulation. The number of convergence criteria


iterationsisalsomodifiedbydoublingtheABAQUSparametersIo,Ir,andIcfromtheir
defaultvalues(seeTable7.1).Iorepresentsthenumberofequilibriumiterationsbeforea
1564H

check is performed to ensure that the magnitudes of the moment and force residual
vectors are decreasing. After Io iterations, if the residuals are not decreasing between
twoconsecutiveequilibriumiterationsthenthelengthoftheincrementstepisreduced
and the equilibrium search is restarted. Ir represents the number of equilibrium

274

iterationsafterwhichthelogarithmicrateofconvergencecheckbegins.Icrepresentsthe
maximumnumberofequilibriumiterationswithinatimeincrementstep. Alinesearch
algorithm is also employed in the STATIC2 model to improve the convergence of the
NewtonRaphson algorithm when nodal force and moment residuals are large. This
algorithmfindsthesolutioncorrectionvectorwhichminimizestheoutofbalanceforces
inthestructuralsystem(ABAQUS2007a).

0.37

STATIC1
STATIC2

P/Py,g

0.9
0.8

0.365

STATIC2 (with line search algorithm)


finds post-peak equilibrium path
before terminating

0.7

P/Py,g

0.6

0.36
0.26

0.27
/t

0.5

0.28

0.4

0.2

0.1
0

Initial imperfection shape


(scale exaggerated)

P
h

0.3

P
h

Displacement control

0.25

0.5

0.75

/t

1.25

1.5

1.75

Figure7.13STATIC1andSTATIC2loaddisplacementcurvesdemonstrateconvergencedifficultiesnearthe
peakload.

Figure7.13comparestheSTATIC1andSTATIC2loaddisplacementcurves.The
156H

STATIC1 model finds the peak load but then terminates due to moment residual
convergenceissuesasitattemptstopredictthefirststepofthepostpeakresponse.The
ABAQUS message (.msg) file for this model states that themoment residuals at nodes
along the loaded edges, along the transverse midline of the plate, and at some nodes

275

near the hole are increasing and convergence is judged unlikely. The solution is
terminated after the automatic time stepping procedure requires a smaller time step
thantheminimumsetinthismodel(1x1020).TheSTATIC2modelwiththelinesearch
algorithm also finds the peak load of the stiffened element and is able to track onto a
postpeak equilibrium path before terminating from the same convergence problems
experienced by the STATIC1 model. The success of the line search algorithm in
identifying a postpeak equilibrium path highlights its potential for solving nonlinear
problems, although a significant increase in computational effort (almost twice the
wallclocktime)wasalsonoted.

1.3.4 NewtonRaphsonwithartificialdamping

173B

The STAB1 and STAB2 models solve the stiffened element problem using a
displacement control Newton Raphson algorithm coupled with the automatic artificial
damping discussed in Section 7.1.2.2. The boundary conditions are modified to those
156H

summarized in Figure 7.8 because of the convergence issues observed with the
1567H

constraintequationsandtransversemidlinerestraints.AsinthecaseofSTATIC1and
STATIC2, the convergence iteration limits Io, Ir, and Ic are doubled from their default
values. In an attempt to alleviate the moment residual convergence issues from

previous runs, the Newton Raphson parameter R n is modified to relax the residual

requirementswhenthesolutionapproachesthepeakload.R nistheallowablelimiton

theratioofthelargestresidualforceormomentatanode(r max)tothelargestchangein
force or moment at a node averaged over each time step increment that has been

276

completed (q). The superscript indicates that R n can be defined for either a

displacement field u or a rotation field . The convergence limit can be written


mathematicallyas:

rmax
Rn q .

(7.3)

ThedefaultforR nof0.005isusedinSTAB1forbothuand fields,whereasinSTAB2


u

R n=0.005andR n=0.100.
1

0.38

STAB1
STAB2

P/Py,g

0.9
0.8

0.34

Highly nonlinear post-peak


equilibrium path found with
STAB1 and STAB2

0.7
0.3
0.25

P/Py,g

0.6

0.3
/t

0.5

0.35

0.4

0.2

0.1
0

Initial imperfection shape


(scale exaggerated)

P
h

0.3

P
h

Displacement control

0.25

0.5

0.75

1.25

/t

1.5

1.75

Figure7.14STAB1andSTAB2loaddisplacementcurvesdemonstrateahighlynonlinearpostpeak
equilibriumpath

The ABAQUS solutions from models STAB1 and STAB2 in Figure 7.14
1568H

demonstrate a highly nonlinear postpeak equilibrium path. Both models are able to
successfullypredictthepeakloadandthenmovetoasecondaryloadpath.Thesolution
terminates because the maximum number of NewtonRaphson iterations is reached.

ThemodificationofthemomentresiduallimitR nfrom0.005to0.100didnotinfluence
thesolutionresults.

277

The STAB3 finite element model employs a uniform loading with the Newton
Raphson algorithm and artificial damping to determine the nonlinear response of the
stiffenedelementwithaslottedhole.TheSTAB3boundaryconditionsarethesameas
thosefortheSTAB1andSTAB2models,wheretheplateedgesareconstrainedtomove
totogetherasrigidbodies(see Figure7.8). Figure7.15comparestheSTAB1andSTAB2
1569H

1570H

(displacement control) to the STAB3 (load control) results and shows that, prior to
yielding,thethreemodelspredictthesameresponse.Differencesintheloadpathsare
observedafteryieldingthough,especiallyintheSTAB3model,whichreachespeakload
and then carries this load with zero stiffness over a large deformation range. This
unstable postpeak behavior results from a complete loss of stiffness as the hole
collapsesunderloadcontrol.Thepeakloadspredictedforthestiffenedelementbythe
displacementcontrolSTAB3modelissevenpercenthigherthantheSTAB1andSTAB2
load control solutions, demonstrating that the peak load is sensitive to the loading
method(uniformloadoruniformdisplacements)inthiscase.

278

STAB1, STAB2
STAB3

0.9
0.4

0.8

0.8
0.35

0.7

0.6

P/Py,g

0.6
0.5

0.3

Higher peak load found with load


control compared to displacement
control (0.395Py,g versus 0.375 Py,g)

0.28 0.32

0.36
Load control solution demonstrates
complete loss of stiffness at peak load
(yielding and plate folding at hole)

0.4

0.4
0.3

P
h

0.2

0.2
0

0.1
0

0.2
0.25

0.5

0.4
0.75

0.6
1

0.8

1.25

1.5

/t

P 1
h

1.75

Figure7.15TheSTAB1andSTAB2models(artificialdamping,displacementcontrol)exhibitasharpdropin
loadasfoldingoftheplateinitiatesnearthehole.TheSTAB3model(artificialdamping,loadcontrol)finds
thecompressiveloadatwhichacompletelossofstiffnessoccurs.

1.4 Influence of a slotted hole on the ultimate strength of a stiffened


element (without geometric imperfections)
98B

The solution controls from the previous section resulting in successful


simulations are now implemented to evaluate the influence of a slotted hole on the
ultimate strength and failure mode of a stiffened element. The loading and boundary
conditions,dimensions,materialproperties,andsolutioncontrolsarethesameasthose
usedfortheSTAB2modeldescribedinSection 7.1.3.4and Table7.1.Initialgeometric
157H

imperfectionsarenotconsidered.

279

1572H

Ultimate Limit State

0.9
0.8

0.8
0.7

Py,net=0.56*Py,g
0.6

P/Py,g

0.6
0.5

Pcr,no hole=0.33*Py,g

0.4

Pcr,hole=0.30*Py,g

0.4
0.3

0.2

0.2
0

0.1
0

Elastic Buckling
0

0.2
0.25

0.4

0.5

0.75

0.6
1

0.8
1.25

/t

1.5

1
1.75

Figure7.16Comparisonofultimatelimitstateandelasticbucklingplatebehavior,initialimperfectionsare
notconsideredintheseresults

Figure7.16demonstratesthattheslottedholereducesthestrengthofthestiffened
1573H

elementfrom1.0Py,g to0.58Py,g,wherePy,gistheresultantcompressiveforceonthe
stiffenedelementtocauseyieldingcalculatedwiththegrosscrosssectionalareaof
the plate. The predicted peak load of the stiffened element with the hole is
consistentwiththeloadatyieldingofthenetsection,Py,net=0.56Py,g.Thisobservation,
thatthestrengthofthestiffenedelementwiththeholeislimitedtoPy,net,highlights
an important consideration in the development of the Direct Strength Method in
Chapter 8. The hole also reduces the axial stiffness of thestiffened element in this
1574H

case, as demonstrated by the change in slope of the linear portion of the load
displacementcurveinFigure7.16.
157H

280

1.5 Influence of geometric imperfection magnitudes on the ultimate


strength of a stiffened element with and without a slotted hole
9B

The ultimate strength of coldformed steel members is sensitive to initial


geometricimperfections.Inthisstudytheinfluenceofimperfectionmagnitudeonthe
ultimatestrengthofstiffenedelementswithandwithoutaslottedholeisevaluated.The
loadingandboundaryconditions,dimensions,materialproperties,andsolutioncontrols
are the same as those used for the STAB2 model discussed in Section 7.1.3.4 and
1576H

summarizedin Table7.1.Theimperfectionfieldisassumedasthefundamentalelastic
157H

bucklingmodepicturedinFigure7.9.Localbuckling(Type1)imperfectionmagnitudes
1578H

correspondingtoP(<d1)=0.25,0.50,0.75,0.95,and0.99fromtheCDFin Figure7.32are
1579H

considered.
Figure7.17and Figure7.18presenttheloaddisplacementresultsforthestiffened
1580H

158H

element without and with the hole and demonstrate that increasing imperfection
magnitudesreducespeakloadandchangepostpeakresponse.Theelasticstiffnessis
alsosoftened,whichcanbeobservedbycomparingtheimperfectionresultstothelinear
slopeoftheloaddisplacementcurvewithoutimperfections.Thissofteningresultsfrom
theinitialoutofplanedeformationswhichengagemoreofthebendingstiffnessofthe
plate and less of the axial stiffness as the plate is compressed. Outofplane
deformations (such as initial imperfections) increase the magnitude of the geometric
stiffness matrix which negates the initial elastic stiffness of the undeformed system.
Imperfections are observed to decrease strength but increase ductility of the stiffened
elementwithandwithoutahole.Theloaddisplacementresultsalsohighlightthatthe

281

holereducestheductilityofthestiffenedelement,whichisconsistentwiththecolumn
experimentresultsinChapter5.
1582H

1
0.9

no imperfections
d1/t=0.14

0.8

d1/t=0.34
d1/t=0.66

0.7

d1/t=1.35
d1/t=3.85

P/Py,g

0.6
0.5
0.4
0.3

P
h

0.2
0.1
0

Displacement
Control

0.25

0.5

0.75

P
h

1.25

1.5

1.75

/t

Figure7.17Loaddisplacementsensitivitytoimperfectionmagnitudeforaplatewithoutahole
1
0.9

no imperfections
d1/t=0.14

0.8

d1/t=0.34
d1/t=0.66

0.7

d1/t=1.35

P/Py,g

0.6

d1/t=3.85

0.5
0.4

P
h

0.3
0.2

0.1
0

P
h

Displacement control

0.25

0.5

0.75

/t

1.25

1.5

1.75

Figure7.18Loaddisplacementsensitivitytoimperfectionmagnitudeforaplatewithaslottedhole

282

1.6

Determination of unstiffened element effective width using


nonlinear finite element modeling
10B

The effective width method provides an approximation to the complex non


uniform stress distribution in a thin buckled plate under compression. Initially
presentedinthe1930sbyvonKarmanandextendedtocoldformedsteelmembersby
Winterinthe1940s,themethodaccountsforthereductioninloadcarryingcapacityof
astiffenedelement(vonKarmanetal.1932;Winter1947).Theinabilityofthecenterof
theplatetocarrycompressiveloadiscausedbyoutofplanedeformationsintheshape
ofthefundamentalelasticbucklingmode.Thesedeformationsreducetheaxialstiffness,
concentratingthecompressiveforceattheedgesofaplate.Theultimateloadisreached
whentheseedgestresses,carriedbytheeffectivewidth,exceedtheyieldstressofthe
plate material. The effective width concept is the basis of most coldformed steel
designcodesaroundtheworldtoday.
In this study, a nonlinear finite element model is employed to calculate the
longitudinal stress distribution at failure for a stiffened element with and without a
slottedhole.Thedistributionofstressesforbothcasesiscompared,andthevariationin
effective width along the length of the stiffened element is determined. The stiffened
element is modeled with the same loading and boundary conditions, dimensions,
materialproperties,andsolutioncontrolsasthoseusedfortheSTAB2modeldiscussed
in Section 7.1.3.4 and described in Table 7.1. The initial imperfection geometry
1583H

1584H

correspondstothefundamentalelasticbucklingmodeoftheplatewithouttheholeas
described in Figure 7.9. d1/t=0.34 is used to scale the initial imperfection field of the
158H

283

plate,whichcorrespondstoaprobabilityofoccurrenceofP(<d1)=0.50asdiscussedin
Section 7.1.1.Theeffectivewidthiscalculatedbyfirstintegratingthelongitudinal(S11)
1586H

membrane stresses at crosssections along the length of the stiffened element and then
dividingtheresultingareasbytheyieldstressofthesteelasshownin Figure7.19.The
1587H

membranestressesarethelongitudinal(S11)stressesthatoccuratthemidplaneofthe
stiffenedelementasdefinedinFigure7.20.
158H

distribute area (A) to


edges of plate

he/2

A/2

calculate area under


stress curve (A)

A/2

he/2

membrane stress (S11)

yield stress

Figure7.19Calculationofeffectivewidthatacrosssectionalongastiffenedelement

+S11

+S11

Plan view of element

Top
Membrane
stress

Midplane

Membrane
stress

Bottom
Elevation view of element

Figure7.20Definitionoflongitudinal(S11)membranestress

284

Figure 7.21(a) highlights the variation in membrane longitudinal stress (S11)


1589H

occurring at the failure load of the stiffened element. The highest stresses accumulate
along the edges of the plate and decrease toward the center of the plate. The largest
edgestressesoccuratthecrestsofthehalfwaveswherethegreystresscontoursindicate
yieldingoftheplate.Thecorrespondingeffectivewidthispresentedin Figure7.21(b).
1590H

The maximum effective width of 0.51 he/h occurs at the inflection point between half
waves, while the minimum effective width of 0.48 he/h occurs at the wave crests. The
predictedeffectivewidthforthisplateusingSectionB2.1oftheAISIspecificationis0.50
he/h(AISIS1002007).

+S11
S, S11
Mid, (fraction = 0.0)
(Ave. Crit.: 75%)
+5.059e-01
-2.286e+00
-5.078e+00
-7.871e+00
-1.066e+01
-1.345e+01
-1.625e+01
-1.904e+01
-2.183e+01
-2.462e+01
-2.742e+01
-3.021e+01
-3.300e+01
-3.578e+01

+S11

Plan view of element

effective width
average
standard deviation
max
min

he/h
0.51
0.02
0.55
0.48

Elevation

(a) membrane stress in 1 direction (S11)

he/2
(b) variation in effective width along plate

Figure7.21(a)longitudinalmembranestressesand(b)effectivewidthofastiffenedelementatfailure

The failure mode of the stiffened element with the slotted hole is fundamentally

differentthanwithoutthehole.Thestressesin Figure7.22(a)demonstratethatyielding
159H

occurs only at the location of the hole when the peak load is reached. Compressive
stressesarehighestattheedgeoftheplateandthentransitiontotensilestressesatthe

285

faceofthehole.Theeffectivewidthoftheyieldedportionoftheplatein Figure7.22(b)
1592H

islessthanthatfortheplatewithoutthehole,evenwiththebeneficialtensilestressesat
theface.Theaverageeffectivewidthis0.38he/h,whichis25percentlessthanthatofthe
stiffenedelementwithoutthehole.ThepredictedeffectivewidthusingSectionB2.2of
theAISISpecificationis0.30he/h.Theeffectivewidthsofthestiffenedelementwithand
withoutaslottedholearecomparedtogetherinFigure7.23.
1593H

+S11
S, S11
Mid, (fraction = 0.0)
(Ave. Crit.: 75%)
+1.151e+01
+7.804e+00
+4.095e+00
+3.852e-01
-3.324e+00
-7.034e+00
-1.074e+01
-1.445e+01
-1.816e+01
-2.187e+01
-2.558e+01
-2.929e+01
-3.300e+01
-3.615e+01

+S11

Plan view of element

effective width
average
standard deviation
max
min

he/h
0.38
0.03
0.41
0.34

Elevation

(a) membrane stress in 1 direction (S11)

he/2

(b) variation in effective width along plate

Figure7.22(a)longitudinalmembranestressesand(b)effectivewidthofastiffenedelementwithaslotted
holeatfailure

he/2 (plate with hole)

he/2 (plate without a hole)

Figure7.23Effectivewidthcomparisonforaplatewithandwithoutaslottedhole

The longitudinal stresses (S11) in the top and bottom fibers of the stiffened
elementatfailurearedifferentfromthemembranestressesatthemidplane,suggesting

286

that the effective width of a stiffened element actually varies through its thickness.
Figure 7.24 and Figure 7.25 provide a comparison of this variation for a stiffened
1594H

159H

elementwithandwithoutaslottedhole.Itisobservedthataplateismoreeffectiveon
thesurfacewheretheoutofplanedeformationcausescompression.Theeffectivewidth
isreducedwhentensileandcompressivestressesnegateeachother,asshowninthe2D
plotofextremefiberandmembranestressesatarepresentativecrosssectionin Figure
1596H

7.26.
Effective width calculated with
longitudinal stresses (S11) at top,
midplane, and bottom of the plate

Top of plate

Midplane of plate

Bottom of plate

Figure7.24Throughthethicknessvariationofeffectivewidthofaplatewithoutahole

Effective Width calculated with


longitudinal stresses (S11) at top,
midplane, and bottom of the plate

Top of plate

Midplane of plate

Bottom of plate

Figure7.25Throughthethicknessvariationofeffectivewidthofaplatewithaslottedhole

287

Longitudinal (S11) stress


variation across width of plate

1
top of plate
midplane of plate
bottom of plate

0.9
0.8
0.7

x/h

0.6
0.5

Top of plate is
fully effective

Stress
distribution used
to calculate
code-based
effective width

0.4
0.3

Tension and compression


stresses counteract each other
when calculating effective width
at the bottom of the plate

0.2
0.1
0
-1.5

Compression

-1

-0.5

Tension

0
fplate/fy

0.5

SECTION A-A

1.5

Figure7.26Throughthicknessvariationinlongitudinal(S11)stressesinaplateatfailure

7.2 Nonlinear finite element modeling of columns


withholes
47B

A more extensive study of ABAQUS nonlinear finite element capabilities of cold

formed steel columns with holes is now presented. Simulation to collapse of the 24
column experiments described in Chapter 5 is performed, considering solution
1597H

sensitivity to specific modeling parameters including initial imperfections, residual


stresses and the coldwork of forming, nonlinear material modeling, and column
boundary conditions. A modeling protocol is developed which produces results
consistent with column experiments. This modeling tool is employed to explore the

288

validity of proposed Direct Strength Method equations for coldformed steel members
withholespresentedinChapter8.
1598H

2.1 Modeling protocol development


10B

2.1.1 Modeldimensionsandfiniteelementmeshing
174B

The collapse behavior of the 24 column specimens is simulated with the general

purposefiniteelementprogramABAQUS(ABAQUS2007a).Allcolumnsaremodeled
withS9R5reducedintegrationninenodethinshellelements(seeSection 2.1fordetails
159H

ontheS9R5element).Thefiniteelementmeshforeachspecimeniscreatedwithcustom
Matlab code developed by the author (see Appendix A); the mesh is consistent with
160H

S9R5 meshing guidelines summarized in Section 2.4. The centerline Csection


160H

dimensionsinputintoABAQUSarecalculatedusingtheouttooutdimensionsofeach
columnspecimenprovidedinTable5.3.Thecrosssectioncorneranglesareassumedas
1602H

right angles (even though they were measured to be off of 90 degrees, see Table 5.4)
1603H

since the distortional imperfection magnitudes obtained in Section 7.2.1.5 are derived
1604H

based on a nominal crosssection with 90 degree corners*. The average base metal
thicknessforeachspecimen(i.e.,theaverageoftbare,w,tbare,f1,andtbare,f2from Table5.5)and
1605H

column length L from Table 5.6 are used to construct the ABAQUS models, as are the
160H

locationoftheslottedwebholesrelativetothecenterlineofthewebprovidedin Table
1607H

5.8.
* The measured flangeweb and weblip angles were not considered because of initial difficulties matching the
experiment results to the simulations. To resolve these difficulties, a simplified model with nominal dimensions was
implemented.(Modelingwithplasticityattheproportionallimitwasfoundtobethetruecauseofthediscrepancies,see
Section 7.2.1.4.)Considerationoftheactualcrosssectiondimensions,includingtheflaredwebflangecornersmeasured
intheexperiments,iswarrantedandisanimportantpointoffuturestudy.
1608H

289

2.1.2 Boundaryconditionsandapplicationofloading
175B

The specimen boundary conditions in ABAQUS are defined to simulate the

experiment boundary conditions as shown in Figure 7.27. The nodes on the loaded
1609H

column face are coupled together in the direction of loading (1 direction) with an
ABAQUSpinnedrigidbodyconstraint.Thisconstraintensuresthatallnodesonthe
loaded face of the column translate together, while the rotational degrees of freedom
remainindependent(asinthecaseofplatenbearing).Atotalimposeddisplacementof
0.20inchesisappliedtothereferencenodeoftheABAQUSrigidbodyoveraseriesof
steps (see Section 7.2.1.3) to simulate the displacement control loading applied by the
160H

bottom platen during the experiment. Frictioncontact boundary conditions were also
evaluated in ABAQUS as described in Appendix J although their influence on the
16H

ultimatestrengthofthecolumnspecimenswasdeterminedtobeminimal.
Nodes bearing on top platen
constrained in 1, 2 and 3

1
2
3

ABAQUS pinned rigid


body reference node
constrained in 2 to 6
directions, ensures that all
nodes on loaded surface
move together in 1 direction

Apply uniform displacement in ABAQUS to


simulate displacement control of experiments

Figure7.27ABAQUSboundaryconditionssimulatingcolumnexperiments

290

2.1.3 Nonlinearsolutionmethod
176B

ThemodifiedRiksmethod(*STATIC,RIKS)isemployedasthesolutionalgorithmin

this study. The preliminary nonlinear finite element studies on stiffened elements
demonstrated that the modified Riks method was able to capture the complete load
deformation response when imposed displacements are used to load the member (see
Figure7.12).ABAQUSautomatictimesteppingwasenabled,withthesuggestedinitial
162H

stepsizesetto0.005,themaximumstepsizelimitedto0.01,andthemaximumnumber
ofincrementsequalto300allinputbytheuser.

2.1.4 Materialmodeling
17B

Steel yielding and plasticity is simulated in ABAQUS using a classical metal

plasticityapproachwithisotropichardening.AMisesyieldsurfaceisdefinedwiththe
true stress and true plastic strain obtained from uniaxial tensile coupon tests for each
specimen.Threestressstraincurves(westflange,eastflange,andweb)wereobtained
for each specimen (see Section 5.2.5). The experimentally obtained engineering stress
163H

straincurvesareconvertedtotruestressandstrainandthenaveragedpointbypointto
produceayieldstress,proportionallimit,andtruestressstraincurveforeachspecimen.
(The true plastic strains and associated stresses are input into ABAQUS with the
*PLASTIC command.) For Mises stresses below the yieldstress, linear elastic material
behaviorisassumedwhereE=29500ksiand=0.3.

291

Preliminary nonlinear modeling efforts for this study determined that including

plastic strains starting at the proportional limit resulted in ABAQUS simulation


predictions that were as much as 25% lower than the column tested strengths. An
example of the average true stressstrain curves with plasticity starting at the
proportional limit are provided in Figure 7.28a for specimen 362124NH and Figure
164H

165H

7.29aforspecimen600124NH(averagetruestressstraincurvesareprovidedforall24
columnspecimensinAppendixH).Otherresearchershaveobtainedsimulationresults
16H

consistentwithexperimentsbyassumingthatplasticstrainsinitiateonlyaftertheyield
stress (determined with the 0.2% offset method) is reached (Schafer 1997; Yu 2005;
Schafer et al. 2006) a material modeling approach that proved to be successful at
predictingthecolumnexperimentpeakloadsforthisstudyalso.Thetruestressstrain
curves in Appendix H were therefore modified to ensure that plasticity initiates in
167H

ABAQUS only after the yield stress is reached for the gradually yielding stressstrain
curves(362S16233specimens,seeFigure7.28b)andattheinitiationoftheyieldplateau
168H

for the sharpyielding curves (600S16233 specimens, see Figure 7.29b). Figure 7.30
169H

1620H

demonstratesthedisparitybetweentheexperimentandFEsimulationloaddeformation
responseforcolumnspecimen600124NHwhenplasticityinitiatesattheproportional
limit.Thereasonsforthisdisparityareunclearandwarrantfuturestudy.Additional
researchworkisplannedtostudythedetailsofmetalplasticityinABAQUSbyloading
asinglefiniteelementtofailureintension.

292

100

100

90

90

80

80

true stress, ksi

ksi

60
50
40

(a)

30
20
10
0

70

true stress

true stress, ksi

true plastic
strain

70

0.02

0
0.001
0.002
0.007
0.012
0.017
0.027
0.037
0.047
0.057
0.067

33.1
46.1
51.9
60.3
64.9
68.4
74.0
78.1
81.3
83.8
86.2

true plastic
true stress
strain
ksi
0
55.1
0.003
60.3
0.008
64.9
0.013
68.4
0.023
74.0
0.033
78.1
0.043
81.3
0.053
83.8
0.063
86.2

60
50
40

(b)

30
20
10

0.04
0.06
0.08
true plastic strain

0.1

0.12

0.02

0.04
0.06
0.08
true plastic strain

0.1

0.12

Figure7.28ABAQUSplasticstraincurveforspecimen362124NHassuming(a)plasticityinitiatesatthe
proportionallimitand(b)plasticityinitiatesat0.2%offsetyieldstress

100

100

90

90

80

80

true stress, ksi

60
50
40
30

(a)

20
10
0

0.02

true stress

0
0.001
0.002
0.007
0.012
0.017
0.027
0.037
0.047
0.057
0.067
0.077
0.087
0.097
0.107

0.04
0.06
0.08
true plastic strain

70

ksi
37.5
54.7
58.3
60.0
61.5
64.0
70.2
74.4
77.5
80.0
81.9
83.5
84.9
86.1
87.2

true stress, ksi

true plastic
strain

70

60
50
40
30

(b)

20
10
0.1

0.12

0.02

true plastic
true stress
strain
ksi
0
58.3
0.005
60.0
0.01
61.5
0.015
64.0
0.025
70.2
0.035
74.4
0.045
77.5
0.055
80.0
0.065
81.9
0.075
83.5
0.085
84.9
0.095
86.1
0.105
87.2

0.04
0.06
0.08
true plastic strain

0.1

0.12

Figure7.29ABAQUSplasticstraincurveforspecimen600124NHassuming(a)plasticityinitiatesatthe
proportionallimitand(b)plasticityinitiatesatthebeginningoftheyieldplateau(refertoAppendixHfor
thedetailsonthedevelopmentofthiscurve).
162H

293

14
plasticity at proportional limit
plasticity at yield stress
experiment

12

axial load, kips

10

0
0

0.02

0.04

0.06

0.08
0.1
0.12 0.14
axial displacement, in.

0.16

0.18

0.2

Figure7.30InfluenceofABAQUSmaterialmodelontheloaddeformationresponseofspecimen600124
NH(workthisfigurewithFigure7.29)
162H

2.1.5 Initialgeometricimperfections
178B

The ultimate strength and failure mechanisms of coldformed steel columns are

sensitivetoinitialgeometricimperfections,asdemonstratedinthepreliminarystudies
on stiffened elements in Section 7.1.5. In this study, the sympathetic local (L) and
1623H

distortional (D) elastic buckling modes are obtained with eigenbuckling analyses for
each column specimen and imposed on the nominal geometry in each finite element
model. (An ABAQUS .fil file is created for each eigenbuckling analysis which is then
calledfromthenonlinear.inpfilewiththe*IMPERFECTIONcommand).Theboundary
conditions at both specimen ends are assumed to be warping free when obtaining the
imperfection shapes (see Figure 4.2 for definition) to ensure consistency with CUFSM
1624H

boundary conditions. Specimens with and without holes are modeled with the same
elasticbucklingimperfectionshapesbyfillingtheholeswithadditionalfiniteelements

294

asshownin Figure7.31.Thisprocedureensuresthattheloaddisplacementbehaviorof
1625H

boththeholeandnoholespecimensarecomparedonequivalentbasis(bothwillhave
the no holeL and D imperfection shapes). Filling in the holes is necessary (instead of
eliminating them completely) because it preserves the nodal numbering and geometry
of thespecimenswithholes, making it convenient to superimpose the L and Dmodes
ontotheinitialnodalgeometryinABAQUS.

Hole is filled in with


S9R5 elements to
produce no hole local
buckling shape

Type1 imperfection (L)

Type 2 imperfection (D)

Figure7.31SlottedholesarefilledwithS9R5elementstoobtainnoholeimperfectionshapes

The magnitudes of the L and D imperfections are determined with the same

probabilistictreatmentusedforthestiffenedelementstudiesinSection 7.1(Schaferand
162H

Pekz 1998). Finite element simulations with L and D imperfection magnitudes


correspondingtothe25thand75thpercentilesoftheCDFinFigure7.32areperformedfor
1627H

eachspecimentoobtainarangeofsimulatedloaddisplacementresponsestocompare
to experiment results. FE simulations are also performed using the L and D elastic

295

bucklingmodeshapesandimperfectionmagnitudesmeasureddirectlyfromthecolumn
specimens.Inthiscasethelocalimperfectionmagnitudeforeachspecimenistakenas
the maximum deviation from the average web elevation as reported in Table 5.9. The
1628H

distortional imperfection magnitude for each specimen is determined by finding the


largest measured angular deviation from 90 degrees along each specimen and
calculatingtheassociatedflangelipdisplacementasshownin Figure7.33.TheType1
1629H

imperfectionmagnitudesmeasuredintheexperimentsareoften2to3timeslargerthan
the 75th percentile CDF magnitudes as shown in Table 7.2. The Type 2 imperfection
1630H

magnitudes for the 362S16233 specimens also are 2 to 3 times larger than the 75th
percentileCDFmagnitudes,primarilybecausethesespecimenstendedtoopenupatthe
sawnends(i.e.,flangewebanglesincreasedabove90degrees)whentheyweresawcut
from full stud lengths. Other researchers have studied this observed change in cross
section after sawcutting (Wang et al. 2006). The 600S16233 specimens were less
sensitive to this sawcutting effect, resulting in measured distortional imperfection
magnitudesconsistentwiththe75thpercentileoftheimperfectionCDF.

1.00
Type 1
Type 2

Type 1

Type 2

(L)

(D)

Probability (X < x)

0.80

0.60

P(X < x)
0.25
0.50
0.75
0.95
0.99

0.40
0.20

Type 1
d/t
0.14
0.34
0.66
1.35
3.87
0.50
0.66

Type 2
d/t
0.64
0.94
1.55
3.44
4.47
1.29
1.07

0.00
0.0

1.0

2.0

3.0

4.0

5.0

d/t
CDF of Maximum Imperfection

296

6.0

7.0

Figure7.32LandDimperfectionmagnitudesdescribedwithaCDF(SchaferandPekz1998)

Deviation from reference


cross section (measured at
X=6, 18 in. for the the short
specimens and X=12, 18, 30
and 36 in. for the 48 in. long
specimens).

D = max(Bi sin (i

where i=1 or 2

Nominal cross-section
1

B1

B2
90 degrees

90 degrees

Figure7.33Methodformeasuringdistortionalimperfectionmagnitudesfromexperiments

Table7.2Localanddistortionalimperfectionmagnitudes
Specimen
362-1-24-NH
362-2-24-NH
362-3-24-NH
362-1-24-H
362-2-24-H
362-3-24-H
362-1-48-NH
362-2-48-NH
362-3-48-NH
362-1-48-H
362-2-48-H
362-3-48-H
600-1-24-NH
600-2-24-NH
600-3-24-NH
600-1-24-H
600-2-24-H
600-3-24-H
600-1-48-NH
600-2-48-NH
600-3-48-NH
600-1-48-H
600-2-48-H
600-3-48-H

Type 1 Imperfection Magnitude (L)

Type 2 Imperfection Magnitude (D)

25% CDF
0.005
0.005
0.005
0.005
0.005
0.006
0.005
0.006
0.005
0.005
0.005
0.006
0.006
0.006
0.006
0.006
0.006
0.006
0.006
0.006
0.006
0.006
0.006
0.006

25% CDF
0.025
0.025
0.025
0.025
0.025
0.025
0.025
0.025
0.025
0.025
0.025
0.026
0.028
0.028
0.028
0.027
0.026
0.028
0.028
0.028
0.028
0.027
0.027
0.028

75% CDF Measured


0.025
0.038
0.025
0.054
0.025
0.036
0.026
0.052
0.025
0.058
0.026
0.044
0.026
0.071
0.026
0.080
0.026
0.057
0.026
0.084
0.026
0.066
0.026
0.050
0.029
0.061
0.029
0.075
0.029
0.096
0.028
0.062
0.027
0.089
0.028
0.087
0.029
0.071
0.028
0.077
0.029
0.073
0.028
0.095
0.028
0.049
0.028
0.068

75% CDF Measured


0.060
0.200
0.060
0.174
0.060
0.205
0.061
0.195
0.059
0.159
0.061
0.140
0.061
0.168
0.061
0.163
0.060
0.184
0.061
0.161
0.061
0.174
0.062
0.156
0.068
0.106
0.068
0.114
0.068
0.114
0.065
0.064
0.064
0.124
0.067
0.102
0.067
0.090
0.067
0.077
0.067
0.074
0.066
0.084
0.067
0.062
0.067
0.099

The initial outofstraightness of each column specimen was measured in the MTS

machineunderasmallpreloadbeforethestartofeachtest.Thisglobalimperfectionis
also superimposed on the nodal geometry for each specimen finite element model as

297

showninFigure7.34.Themagnitudeoftheglobalimperfection,g,isprovidedinTable
163H

1632H

7.3.
Loading Line

g (+ shown)

Specimen COG (Typ.)

Section a-a

Figure7.34DefinitionofoutofstraightnessimperfectionsimplementedinABAQUS

Table7.3Outofstraightnessimperfectionmagnitudes
Specimen
362-1-24-NH
362-2-24-NH
362-3-24-NH
362-1-24-H
362-2-24-H
362-3-24-H
362-1-48-NH
362-2-48-NH
362-3-48-NH
362-1-48-H
362-2-48-H
362-3-48-H
600-1-24-NH
600-2-24-NH
600-3-24-NH
600-1-24-H
600-2-24-H
600-3-24-H
600-1-48-NH
600-2-48-NH
600-3-48-NH
600-1-48-H
600-2-48-H
600-3-48-H

298

g
in.
-0.024
0.004
0.038
-0.012
0.034
-0.023
0.047
-0.028
0.012
0.066
0.013
-0.003
-0.063
-0.141
0.063
-0.078
0.076
0.069
-0.036
-0.087
-0.049
-0.098
0.072
0.020

2.1.6 Residualstressesandequivalentplasticstrains
179B

Chapter6describesageneralmethodforpredictingthethroughthicknessresidual
163H

stressesandstrainsincoldformedsteelmemberswhichcanthenbereadilyinputinto
ABAQUS.The prediction method assumes that residual stresses and plastic strains
occuroverthefullcrosssectionfromcoiling,uncoiling,andflatteningofthesheetcoil.
The coiling residual stresses are largest when the sheet thickness t is large (>0.068 in.)
and the yield stress is low (<40 ksi). The predicted coiling, uncoiling, and flattening
residual stresses (and plastic strains) are zero in this study because the column
specimens have a relatively low sheet thickness (~0.040 in.) and high yield stress (~ 60
ksi).

Residual stresses and plastic strains from the rollforming of the crosssection are

consideredinthisstudy.ThesestressesareappliedinABAQUSwiththeelementlocal
coordinate system shown in Figure 7.35 starting from section point 1 (SNEG). The
1634H

transverse residual stress distribution (2direction) is provided in Figure 7.36 and the
1635H

longitudinaldistribution(1direction)inFigure7.37asafunctionofyieldstressyield(yield
163H

islistedin Table5.13foreachspecimen).PlasticstrainsareinputintoABAQUSinvon
1637H

Mises space and therefore only plastic strain magnitudes are required, not a specific
direction.Thehighestinitialstrainsoccurattheinnerandoutersurfacesofthecorners
asshowninFigure7.38.pisapproximatedusingtheprocedureoutlinedinFigure6.17.
1638H

1639H

299

Element normal
SPOS
1
SNEG
2

Figure7.35ABAQUSelementlocalcoordinatesystemforusewithresidualstressdefinitions

SPOS
-0.50yield

-yield

+yield

+0.50yield
SNEG

Figure7.36Transverseresidualstressdistributionappliedatthecornersofthecrosssection

SPOS
+0.05yield

+0.50yield

-0.50yield

-0.05yield
SNEG
Figure7.37Longitudinalresidualstressdistributionappliedatthecornersofthecrosssection

300

SPOS

SNEG

Figure7.38Equivalentplasticstraindistributionatthecornersofthecrosssection

The transverse residual stress distribution has the special property that it is self

equilibratingforbothmomentandaxialforce,i.e.thetotalforceandmomentthrough
thethicknessiszero.Thisselfequilibratingcharacteristicensuresthatnodeformation
(orredistributionofstress)willoccurinABAQUSintheinitialstate.Thelongitudinal
stress distribution is selfequilibrating for axial force but not for moment. The
deformationsassociatedwiththisoutofbalancebendingmomentareinfinitesimaland
verysmallredistributionsinstressareobserved(0.1ksi)intheinitialstate.

Thenumberofelementsectionpointsthroughthethicknessdictatestheaccuracyof

the residual stress distribution. If only a small number of section points are used, the
discontinuity in stress at the middle thickness cannot be modeled accurately and
excessive transverse deformations of the crosssection will occur. Figure 7.39
1640H

demonstratesthedecreaseinunbalancedthroughthicknesstransversemoment,MUB,as
the number of section points increase (sheet thickness is assumed as t=0.040 in. and

yield=60 ksi). As the number of section points decrease, the residual stress approaches
0.50My,whereMyistheyieldmomentofthesheetsteelperunitwidthdefinedas:

301

t2
M y = yield
6

(7.4)

55 section points are used in the specimen finite element models for this study as a
compromise between model accuracy and computational cost. ABAQUS limits the
maximumnumberofsectionpointsto250fortheS9R5element(ABAQUS2007b).

0.5
0.45
0.4
0.35
MUB/My

0.3
0.25
0.2
0.15
0.1
0.05
0

20
40
60
80
# of element through-thickness section points

100

Figure7.39Influenceofsectionpointsontheunbalancedmoment(accuracy)ofthetransverseresidual
stressdistributionasimplementedinABAQUS

2.2 Modeling protocol validation


102B

2.2.1 Ultimatestrengthandfailuremechanisms
180B

The nonlinear finite element protocol presented in Section 7.2.1 is demonstrated to


164H

be a viable, conservative predictor of peak load when compared to the experiment


resultsin Table7.4.ThemeanoftheexperimentalpeakloadPtesttoABAQUSpeakload
1642H

302

PABAQUSratiosare1.03(25thpercentileimperfectionCDF),1.05(75thpercentileimperfection
CDF), and 1.11 (measured imperfections). In a few cases (and always with specimens
withholes),ABAQUSwasnotabletoobtainthepeakload,eitherbecausethemodified
Riks solution algorithm reversed the direction of the applied load (similar to that
observedin Figure7.12forstiffenedelements)orbecausetheABAQUScouldnotfind
1643H

equilibriumandterminatedthesimulation.Asimperfectionmagnitudesincreased,the
modifiedRikssolutionalgorithmwasmoresuccessfulatreachingpeakload.Thistrend
is hypothesized to occur because for small imperfection magnitudes a specific
deformationpatternisnotestablishedandmanyequilibriumpathsexistnearpeakload,
whereas for larger imperfection magnitudes a dominate deformation shape and
equilibrium path are defined early in the simulation. Nonlinear FE loaddisplacement
behavioris providedforarepresentativesampleofspecimensin Figure7.40to Figure
164H

1645H

7.47,includingtheloaddisplacementcurvesanddeformedshapeatcollapse(compare
these simulated shapes to the pictures of experiments in Appendix F). FE simulation
164H

loaddisplacementcurvesareprovidedforallspecimensinAppendixI.
1647H

Table7.4ComparisonofnonlinearFEsimulationpeakloadstoexperiments

303

Specimen

25th percentile
imperfection CDF
PABAQUS Ptest/PABAQUS
kips
10.26
1.02
10.13
1.04
10.21
0.99
9.22
1.09
8.83
1.18
9.19
1.08
9.48
0.96
9.40
1.01
9.26
1.02
8.97
1.00
8.91
1.03
8.58
1.09
12.14
0.98
12.10
0.99
12.10
1.01
DNC
--11.10
1.05
DNC
--11.27
0.99
11.27
1.02
11.37
0.99
DNC
--DNC
--DNC
--1.03
0.05

Ptest
kips
10.48
10.51
10.15
10.00
10.38
9.94
9.09
9.49
9.48
8.95
9.18
9.37
11.93
11.95
12.24
12.14
11.62
11.79
11.15
11.44
11.29
11.16
11.70
11.16

362-1-24-NH
362-2-24-NH
362-3-24-NH
362-1-24-H
362-2-24-H
362-3-24-H
362-1-48-NH
362-2-48-NH
362-3-48-NH
362-1-48-H
362-2-48-H
362-3-48-H
600-1-24-NH
600-2-24-NH
600-3-24-NH
600-1-24-H
600-2-24-H
600-3-24-H
600-1-48-NH
600-2-48-NH
600-3-48-NH
600-1-48-H
600-2-48-H
600-3-48-H
Average
Standard deviation
DNC

75th percentile
imperfection CDF
PABAQUS Ptest/PABAQUS
kips
9.88
1.06
9.70
1.08
9.85
1.03
9.08
1.10
8.70
1.19
9.11
1.09
9.34
0.97
9.27
1.02
8.89
1.07
8.73
1.02
8.63
1.06
DNC
--12.03
0.99
12.01
1.00
11.99
1.02
11.63
1.04
11.08
1.05
11.76
1.00
11.14
1.00
11.39
1.00
11.18
1.01
10.22
1.09
DNC
--10.35
1.078
1.05
0.05

Measured imperfections
PABAQUS
kips
8.72
8.82
8.69
8.48
8.27
8.78
7.76
8.36
7.44
8.30
8.26
ED
11.83
11.74
11.64
11.45
10.82
11.49
11.32
11.30
11.04
ED
10.17
10.30

Ptest/PABAQUS
1.20
1.19
1.17
1.18
1.26
1.13
1.17
1.14
1.28
1.08
1.11
--1.01
1.02
1.05
1.06
1.07
1.03
0.98
1.01
1.02
--1.15
1.08
1.11
0.08

Did Not Complete, Abaqus terminated before finding the peak load

ED Excessive distortion - Abaqus error, imperfection magnitude causes element distortional

Theinitialelasticslopeofthe25%CDFand75%CDFFEloaddisplacementcurves

are consistent with experimental results as shown in Figure 7.40 to Figure 7.47,
1648H

1649H

demonstratingthattheelasticmaterialmodelingassumptionsandspecimendimensions
areconsistentwiththeexperiments.Theinitialslopeoftheloaddisplacementcurveis
also sensitive to imperfection magnitudes, and therefore the similarities between
experiment and the FE results confirm the assumption that the 25th and 75th percentile
imperfectionmagnitudesintheFEsimulationsproducephysicallyrealisticresults.This
is contrary to the FE simulations with measured imperfections for the 362S16233
specimens(forexample,see Figure7.40),wheretheinitialloaddisplacementslopeand
1650H

peak load are 15% to 30% less than the experimental results (see Table 7.4 and Figure
165H

1652H

7.40). The FE simulations for the 600S16233 specimens are much less sensitive to
imperfectionmagnitudes(forexample,seeFigure7.44).Themaximumdifferenceintest
1653H

to predicted ratio between the three imperfection levels (25% CDF, 75% CDF, and
measured)inTable7.4forthe600S16233specimensis3%.
1654H

304

The postpeak ductility of the column specimens is often underpredicted in the

ABAQUSnonlinearfiniteelementmodels.Thecollapsemechanismofacolumndictates
itsductilityandinsomecasesitspeakload.Forexample,outwarddistortionalbuckling
has been shown to produce lower column strengths than inward distortional buckling
(SilvestreandCamotim2005).ThisobservationcouldexplainwhytheFEsimulationsof
the362S16233specimenswithholes(whichexhibitoutwarddistortionalbuckling)have
alowerpeakloadandductilitythantheexperimentresults(allthreespecimensexhibit
inward distortional buckling, see Appendix F). Another factor influencing column
165H

ductility may be the ABAQUS material modeling effect discussed in Section 7.2.1.4.
165H

Whenplasticityisconsideredattheproportionallimit(see Figure7.30)thepeakofthe
1657H

loaddisplacement curve is flattened which ismore consistent with experiment results.


Thesehypothesesmotivateimportantfutureworktobetterunderstandmetalplasticity
andmaterialmodelinginABAQUSandalsotheinfluenceofimperfectionshapesonFE
columnductilityandstrengthpredictions.

305

14
Experiment
FE, 25% Imperfection CDF
FE, 75% imperfection CDF
FE, measured imperfections

12

axial load, kips

10

0
0

0.02

0.04

0.06

0.08
0.1
0.12 0.14
axial displacement, in.

0.16

0.18

0.2

75% Imperfection CDF

Figure7.40Loaddisplacementresponseofspecimen362124NH

14
Experiment
FE, 25% Imperfection CDF
FE, 75% imperfection CDF
FE, measured imperfections

12

axial load, kips

10

0
0

0.02

0.04

0.06

0.08
0.1
0.12 0.14
axial displacement, in.

0.16

0.18

0.2

75% Imperfection CDF

Figure7.41Loaddisplacementresponseofspecimen362124H

306

14
Experiment
FE, 25% Imperfection CDF
FE, 75% imperfection CDF
FE, measured imperfections

12

axial load, kips

10

0
0

0.02

0.04

0.06

0.08
0.1
0.12 0.14
axial displacement, in.

0.16

0.18

0.2

75% Imperfection CDF

Figure7.42Loaddisplacementresponseofspecimen362148NH

14
Experiment
FE, 25% Imperfection CDF
FE, 75% imperfection CDF
FE, measured imperfections

12

axial load, kips

10

0
0

0.02

0.04

0.06

0.08
0.1
0.12 0.14
axial displacement, in.

0.16

0.18

0.2

75% Imperfection CDF

Figure7.43Loaddisplacementresponseofspecimen362148H

307

14
Experiment
FE, 25% Imperfection CDF
FE, 75% imperfection CDF
FE, measured imperfections

12

axial load, kips

10

0
0

0.02

0.04

0.06

0.08
0.1
0.12 0.14
axial displacement, in.

0.16

0.18

0.2

75% Imperfection CDF

Figure7.44Loaddisplacementresponseofspecimen600124NH

14
Experiment
FE, 25% Imperfection CDF
FE, 75% imperfection CDF
FE, measured imperfections

12

axial load, kips

10

0
0

0.02

0.04

0.06

0.08
0.1
0.12 0.14
axial displacement, in.

0.16

0.18

0.2

75% Imperfection CDF

Figure7.45Loaddisplacementresponseofspecimen600224H

308

14
Experiment
FE, 25% Imperfection CDF
FE, 75% imperfection CDF
FE, measured imperfections

12

axial load, kips

10

0
0

0.02

0.04

0.06

0.08
0.1
0.12 0.14
axial displacement, in.

0.16

0.18

0.2

75% Imperfection CDF

Figure7.46Loaddisplacementresponseofspecimen600148NH

14
Experiment
FE, 25% Imperfection CDF
FE, 75% imperfection CDF
FE, measured imperfections

12

axial load, kips

10

0
0

0.02

0.04

0.06

0.08
0.1
0.12 0.14
axial displacement, in.

0.16

0.18

0.2

75% Imperfection CDF

Figure7.47Loaddisplacementresponseofspecimen600348H

309

2.2.2 Influenceofresidualstressesandinitialplasticstrains
18B

Residualstresses(RS)andinitialplasticstrains(PS)fromthemanufacturingprocess

are approximated with the prediction method in Chapter 6 and then input into
1658H

ABAQUSasdiscussedinSection 7.2.1.6. Figure7.48highlightstheireffectontheload


1659H

160H

deformation response of specimen 600124NH. A small increase in peak load


(approximately 2%) is observed when just initial plastic strains are considered at the
corners, which simulates the increase in apparent yield stress from strain hardening.
Theincreaseinstrengthisminimalbecausetheinfluenceofthestiffenedcornersisoffset
bythelargeproportionofunformedsteel(i.e.,flats)inthecrosssection.Thetransverse
andlongitudinalresidualstressescreatedbycoldbendingofthecrosssectionalsohave
aminimalimpactontheloaddeformationresponseforthisspecimen,primarilybecause
theplasticstrainsatthecornersarehigh(pispredictedtobelargeas0.20atthecorner
outerfibers)whichincreasestheapparentyieldstressinABAQUSandpreventsalossin
stiffnessatthecorners,evenwiththepresenceofthethroughthicknessresidualstresses
inthecolumn.Similarloaddisplacementtrendsarealsoobservedforspecimen3621
24NHasshowninFigure7.49.
16H

Residual stresses and plastic strains are expected to have a larger influence on the

ultimate strength of members with crosssections made from thicker sheet steel, since
coiling and uncoiling of the sheet steel will impart residual stresses and plastic strains
around the entire crosssections (see Figure 6.19). Future research is planned to study
162H

the influence of throughthickness residual stresses and plastic strains on yielding

310

patterns and failure modes of coldformed steel members. The ABAQUS metal
plasticitymodelwithisotropicversuskinematichardeningalsoneedsfurtherstudyto
determine if one is better than the other when considering the influence of residual
stressesandstrains(seeSection6.7foramoredetaileddiscussion).
163H

14
without RS or PS
with RS and PS
experiment

12

axial load, kips

10

0
0

0.02

0.04

0.06

0.08
0.1
0.12 0.14
axial displacement, in.

0.16

0.18

0.2

Figure7.48Influenceofresidualstresses(RS)andplasticstrains(PS)ontheFEloaddisplacementresponse
ofspecimen600124NH

311

14
without RS or PS
with RS and PS
experiment

12

axial load, kips

10

0
0

0.02

0.04

0.06

0.08
0.1
0.12 0.14
axial displacement, in.

0.16

0.18

0.2

Figure7.49Influenceofresidualstresses(RS)andplasticstrains(PS)ontheFEloaddisplacementresponse
ofspecimen362124NH.

312

Chapter 8
The Direct Strength Method for coldformed steel members with holes
7B

Thenonlinearfiniteelementcapabilitydevelopedin Chapter7isnowemployedto
164H

evaluate proposed Direct Strength Method (DSM) formulations for coldformed steel
members with holes. Several hundred coldformed steel columns and beams with
standardSSMAstructuralstudcrosssections(SSMA2001)andwithvaryingwebhole
sizes,shapes,andspacingsaresimulatedtocollapseinABAQUS.Theelasticbuckling
properties of these members (Pcrl, Pcrd, and Pcre for columns and Mcrl, Mcrd, and Mcre for
beams),includingthepresenceoftheholes,areapproximatedwiththeCUFSMelastic
buckling prediction methods developed in Chapter 4. The corresponding ultimate
165H

strengths(obtainedfromtheABAQUSsimulations)aremergedwiththeelasticbuckling
informationintoasimulatedexperimentsdatabasewhichisutilizedtoidentifypotential
modifications to the existing DSM local, distortional, and global failure prediction
curves. Specific DSM options are proposed from these comparisons, which are then

313

comparedtotheexperimentelasticbucklingandtestedstrengthdatabasesin Chapter4
16H

to formalize the final proposed DSM recommendations for coldformed steel members
withholes.

8.1 DSMforcolumnswithholes
48B

1.1 Database of simulated column experiments


103B

Simulated experiments were conducted on 211 Csection columns with evenly

spaced slotted or circular web holes in ABAQUS. Column lengths and crosssections
werespecificallyselectedwithcustomMatlabcodeemployingtheexistingDSMdesign
curves to identify columns predisposed to local, distortional, and global buckling type
failures. The crosssections were chosen from a catalog of 99 industry standard C
sections published by the Steel Stud Manufacturers Association (SSMA 2001). The
nominal outtoout dimensions provided in the SSMA catalog were converted to
centerline dimensions and then constructed in ABAQUS with the meshing procedure
describedinSection 7.2.1.1.Evenlyspacedcircularorslottedwebholeswereplacedin
167H

the columns with hole spacing S (defined in Figure 3.2) varying between 12 and 22
168H

inches.Theholeswerecenteredtransverselyinthewebandtheirdepth,hhole,wasvaried
suchthattheratioofthenetcrosssectionalarea,Anet,tothegrosscrosssectionalarea,Ag,
rangedbetween0.60and1.0.

TheABAQUS boundaryconditionsandapplicationofloading,describedin Figure


169H

8.1,areimplementedtobeconsistentwithCUFSM,i.e.pinnedpinnedandfreetowarp
with a uniform stress applied at the member ends. These boundary conditions were

314

specifically chosen to permit the use of CUFSM simplified elastic buckling methods
whenpredictingtheelasticbucklingbehaviorofcolumnswithholes.(Ifpinnedpinned
warpingfixed end conditions or fixedfixed end conditions were used the elastic
buckling predictions would have required modifications factors, see Eq. (4.8) for an
1670H

example). CUFSMboundaryconditionsrepresentalower boundon memberstrength


arethereforeconsideredconservativeindesign.Consistentnodalloadsareappliedto
simulate the uniform compressive stress at the column ends (see Section 7.2.1.2 for
167H

information on S9R5 consistent nodal loads). The loads (areference load of 1 kip was
appliedateachendinABAQUS)aredistributedoverthefirsttwosetsofcrosssection
nodestoavoidlocalizedfailuresattheloadededges.
End cross-section nodes
restrained in 2 and 3

Node centered in flange at


longitudinal midline restrained in 1
(to prevent rigid body motion)

Consistent nodal loads applied


over two sets of cross-section
nodes to avoid edge failures
End cross-section nodes
restrained in 2 and 3

1
4

5
2

Figure8.1ABAQUSsimulatedcolumnexperimentsboundaryconditionsandapplicationofloading

The ABAQUS simulations were performed with the modified Riks nonlinear

solution algorithm. Automatic time stepping was enabled with a suggested initial arc

315

lengthstepof0.25(theRiksmethodincrementsinunitsofenergy,inthiscasekipin.),a
maximumstepsizeof0.75,andthemaximumnumberofsolutionincrementssetat300.
Metal plasticity was simulated with the material modeling procedure described in
Section7.2.1.4.Theplastictruestressstraincurveforspecimen362148HinAppendix
1672H

1673H

Hwasassumedforallcolumnmodels(butmodifiedsothatplasticitystartsattheyield
stress,seeSection 7.2.1.4),wherethesteelyieldstressFy=58.6ksi.Residualstressesand
1674H

initialplasticstrains,asdiscussedinSection7.2.1.6,werenotconsideredintheABAQUS
1675H

models because their implementation requires further validation and they were not
observed to markedly influence column ultimate strength (see Figure 7.48 and Figure
167H

167H

7.49).

Imperfections were imposed on the initial column geometry in ABAQUS with

custom Matlab code which combines the local, distortional buckling, and global cross
section mode shapes from CUFSM along the column length. Two simulations were
performed for each column, one model with 25% CDF local and distortional
imperfectionmagnitudesandL/2000globalimperfections(whereListhelengthofthe
column) and the other model with 75% CDF local and distortional imperfection
magnitudesandaglobalimperfectionmagnitudeofL/1000(seeSection 7.2.1.5forlocal
1678H

anddistortionalimperfectiondefinitions).

Theglobalimperfectionmagnitudeassumptionsarebasedonhotrolledcolumnout

ofstraightness measurements (Galambos 1998b) because no formal guidelines are


currently available for coldformed steel columns. The use of hotrolled steel column
imperfection magnitudes is consistent with the DSM approach for global buckling

316

controlledfailures.DSMemploysthesameglobaldesigncurveasthatspecifiedbythe
Structural Stability Research Council (SSRC) for hotrolled steel (Galambos 1998b),
thereby indirectly assuming that the influence of hotrolled steel global imperfection
magnitudes also apply to coldformed steel. The global imperfection shape of the
columns in the simulation database was eitherweakaxis flexural buckling or flexural
torsionalbuckling,dependingonthecrosssectiondimensionsandlengthofthecolumn.
Csectionsarenotsymmetricabouttheirweakbendingaxis,andthereforethedirection
of the global imperfection influences the predicted strength when weakaxis flexural
bucklingdefinestheglobalimperfectionshape(e.g.,webincompressionfrombowing
or flange lips in compression from bowing). Simulations with both L/1000 and
L/2000 imperfection magnitudes were performed to capture this strength effect for
weakaxisflexuralbucklingmodeshapes.Globalimperfectionswerenotconsideredfor
columns with L/D18 (i.e., stockier columns with a low sensitivity to global
imperfections),whereDistheouttooutflangewidthofthecolumn.

Thelocal(Pcrl),distortional(Pcrd),andglobal(Pcre)criticalelasticbucklingloadswere

predicted for each column with custom Matlab code based on the CUFSM prediction
methods described in Section 4.2.7. The database of simulated column experiments,
1679H

including crosssection type, column and hole geometry, simulated ultimate strength
(Ptest25 and Ptest75) and critical elastic buckling loads for each column (including the
presenceofholes)isprovidedinAppendixK.
1680H

317

1.2 Distortional buckling study


104B

Agroupof20columnsfromtheSSMAcolumnsimulationdatabasewaschosento

evaluatetheinfluenceoftheratioAnet/Ag onthetestedstrengthofcolumnspredictedto
collapsewithadistortionalfailuremode.Agisthegrosscrosssectionalareaofacolumn
and Anet is the crosssectional area at the location of a hole. In this study the column
length,L,isheldconstantat24in.andthecolumnwidthsrangefrom6in.to12in.The
SSMA crosssections chosen have relatively thick sheet steel (t up to 0.1017 in.) which
preventsalocalbucklingtypefailure.Thewebofeachcolumnhastwocircular holes
wheretheholespacingS=12in(see Figure3.2forthedefinitionofS).Theholedepth
168H

(diameter),hhole,isvariedforeachcolumntoproduceAnet/Agof1.0(noholes),0.9,0.8,0.7,
and 0.6. Refer to Appendix K, Study Type D, for specific crosssection and hole
1682H

geometry information for each column. Figure 8.2 provides an example of an SSMA
1683H

600S25097structuralstudcolumnconsideredinthestudy.
SSMA 800S250-97 structural stud column

Anet/Ag

1.0

0.90

0.80

0.70

0.60

Figure8.2SSMA800S25097structuralstudwithwebholesconsideredintheDSMdistortionalbuckling
study

ThesimulationresultsforAnet/Ag=1.0,0.9,0.8,0.7,and0.6arecomparedtotheDSM

distortionalbucklingpredictioncurveinFigure8.4toFigure8.8.Thecolumnstrengths,
1684H

1685H

Ptest25 and Ptest75, without holes (Anet/Ag=1.0)are consistent with theDSM design curve as

318

shownin Figure8.4a,confirmingtheviabilityofthenonlinearsimulationprotocol.The
168H

mean and standard deviation of the simulated test to predicted ratio is 1.10 and 0.10
respectivelyfor25%CDFlocalanddistortionalimperfections,and1.06and0.13for75%
CDF imperfections (global imperfections are not considered in these stocky columns).
For the columns with holes, the simulated test strengths diverge from the DSM
predictioncurveasdistortionalslenderness, d=(Pyg/Pcrd)0.5,decreasesasshownin Figure
1687H

8.5ato Figure8.8a(Pygisthesquashloadofthecolumncalculatedwiththegrosscross
168H

sectional area Ag). This divergent trend in Ptest with decreasing d can be explained as
follows.When dishigh(i.e.PcrdislowrelativetoPyg),thecolumnstrengthislowerthan
Pygbecausethecollapsemechanismiscontrolledbydistortionalbucklingdeformations.
The presence of a hole may decrease Pcrd (as predicted with the method in Section
4.2.7.2), but the distortional failure mechanismstill dominates in this case. When d is
1689H

low, Pcrd is much higher than Pyg and the column is not as sensitive to distortional
deformation.Instead,thecolumnfailsbyyieldingofthecrosssection.Whenaholeis
added,theyieldingofthecrosssectionoccursatthelocationofthehole(i.e.,atthenet
section) resulting in the collapse of the unstiffened strips adjacent to the hole. This
collapse is accompanied by distortional and global deformations caused by the
reduction in stiffness at the net section. These two column failure mechanisms, a
distortional buckling failure (when d is high) and yielding and collapse of the net
section(whendislow),arecomparedinFigure8.3.
1690H

319

Yielding and collapse of the unstiffened strips adjacent


to the holes accompanied by distortional and global
(weak axis flexural) deformation

Distortional buckling failure

Anet/Ag

1.0

0.90

0.80

0.70

0.60

Ptest25/Pyg

0.60

0.59

0.57

0.54

0.45

1.33

1.36

1.39

1.42

1.45

Figure8.3SSMA800S25097structuralstudfailuremodetransitionfromdistortionalbucklingtoyieldingat
thenetsection

TheobservationsfromthisstudyareusedtoformulateamodifiedDSMdistortional

curve for columns with holes which captures the failure mechanism transition from
yielding at the net crosssection to a distortional type failure mode and limits the
strengthofthecolumntoitssquashloadatthenetsection:

Distortional Buckling
The nominal axial strength, Pnd, for distortional buckling shall be calculated in accordance with the
following:
(a) For d d1
Pnd = Pynet

(caponcolumnstrength)

(b) For d1<dd2


Pnd = P Pynet Pd 2
ynet

d1
d2

(yieldcontroltransition)

(c) For d > d2

Pnd = 1 0.25 crd


Py

where

0 .6

Pcrd
P
y

0 .6

(existingDSMdistortionalcurve)

Py

d1

= 0.561(Pynet Py )

Py Pcrd

d2

= 0.561 14 (Pynet Py )0.4 13

Pd2

= 1 0.25 (1 d 2 )

1.2

)(1

1 .2

d2

Py

Pynet = FyAnet0.6Py
= Column cross-sectional area at the location of hole(s)

Anet
Pcrd

= Critical elastic distortional column buckling load including hole(s)

320

ThemodifiedDSMdistortionalcurveisaddedinFigure8.5btoFigure8.8basAnet/Ag
169H

1692H

decreases,simulatingthetransitionfromtheexistingDSMcurvetothecappedcolumn
strength exhibited by the simulated test data. The linear portion of the modified
prediction curve represents the unstiffened strip distortional collapse mechanism and
thenonlinearportionrepresentsacollapsemechanismdrivenbydistortionalbuckling.
ThisproposedmodificationtotheDSMdistortionalpredictioncurvewillbecompared
againstthecolumnexperimentsdatabasedevelopedinSection4.2.6.1asapartofseveral
1693H

proposedDSMoptionsconsideredlaterinthischapter.
1.4

1.4
DSM (no hole)
FE 25% CDF imperfections
FE 75% CDF imperfections

0.8

0.8

0.6

0.6

0.4

0.4

0.2

0.2

0.5

1.5

2.5

3.5

DSM (no hole)


DSM (proposed, with holes)
FE 25% CDF imperfections
FE 75% CDF imperfections

1.2

Ptest/Py

Ptest/Py

1.2

0.5

distortional slenderness, d=(Py /P crd)0.5

1.5

2.5

3.5

distortional slenderness, d=(Py /P crd)0.5

Figure8.4Comparisonofsimulatedcolumnstrengths(Anet/Ag=1.0)to(a)theexistingDSMdistortional
bucklingdesigncurveandto(b)theproposedDSMdistortionalbucklingcurveforcolumnswithholes
1.4

1.4
DSM (no hole)
FE 25% CDF imperfections
FE 75% CDF imperfections

0.8

0.8

0.6

0.6

0.4

0.4

0.2

0.2

0.5

1.5

2.5

3.5

DSM (no hole)


DSM (proposed, with holes)
FE 25% CDF imperfections
FE 75% CDF imperfections

1.2

Ptest/Py

Ptest/Py

1.2

distortional slenderness, d=(Py /P crd)0.5

0.5

1.5

2.5

distortional slenderness, d=(Py /P crd)0.5

3.5

Figure8.5Comparisonofsimulatedcolumnstrengths(Anet/Ag=0.90)to(a)theexistingDSMdistortional
bucklingdesigncurveandto(b)theproposedDSMdistortionalbucklingcurveforcolumnswithholes

321

1.4

1.4
DSM (no hole)
FE 25% CDF imperfections
FE 75% CDF imperfections

0.8

0.8

0.6

0.6

0.4

0.4

0.2

0.2

0.5

1.5

2.5

3.5

DSM (no hole)


DSM (proposed, with holes)
FE 25% CDF imperfections
FE 75% CDF imperfections

1.2

Ptest/Py

Ptest/Py

1.2

0.5

distortional slenderness, d=(Py /P crd)0.5

1.5

2.5

3.5

distortional slenderness, d=(Py /P crd)0.5

Figure8.6Comparisonofsimulatedcolumnstrengths(Anet/Ag=0.80)to(a)theexistingDSMdistortional
bucklingdesigncurveandto(b)theproposedDSMdistortionalbucklingcurveforcolumnswithholes
1.4

1.4
DSM (no hole)
FE 25% CDF imperfections
FE 75% CDF imperfections

0.8

0.8

0.6

0.6

0.4

0.4

0.2

0.2

0.5

1.5

2.5

3.5

DSM (no hole)


DSM (proposed, with holes)
FE 25% CDF imperfections
FE 75% CDF imperfections

1.2

Ptest/Py

Ptest/Py

1.2

0.5

distortional slenderness, d=(Py /P crd)0.5

1.5

2.5

3.5

distortional slenderness, d=(Py /P crd)0.5

Figure8.7Comparisonofsimulatedcolumnstrengths(Anet/Ag=0.70)to(a)theexistingDSMdistortional
bucklingdesigncurveandto(b)theproposedDSMdistortionalbucklingcurveforcolumnswithholes
1.4

1.4
DSM (no hole)
FE 25% CDF imperfections
FE 75% CDF imperfections

0.8

0.8

0.6

0.6

0.4

0.4

0.2

0.2

0.5

1.5

2.5

3.5

DSM (no hole)


DSM (proposed, with holes)
FE 25% CDF imperfections
FE 75% CDF imperfections

1.2

Ptest/Py

Ptest/Py

1.2

distortional slenderness, d=(Py /P crd)0.5

0.5

1.5

2.5

distortional slenderness, d=(Py /P crd)0.5

3.5

Figure8.8Comparisonofsimulatedcolumnstrengths(Anet/Ag=0.60)to(a)theexistingDSMdistortional
bucklingdesigncurveandto(b)theproposedDSMdistortionalbucklingcurveforcolumnswithholes

322

1.3 Global buckling study


105B

This study compares simulated strengths to DSM predictions of coldformed steel

columnswithholespredictedtoexperienceaglobalfailure.Aglobalfailureistriggered
by yielding for a stocky column and flexural or flexuraltorsional buckling for slender
columns.NomodificationsareproposedtotheDSMglobalbucklingdesigncurvefor
columns with holes, as the influence of holes on short columns will be accounted for
withtheDSMlocalbucklingdesigncurve(seeSection8.1.4).Forexample,whenPne=Pyg,
1694H

Pnl will always be made less than or equal to Pynet, and therefore the nominal column
strength,Pn,willalwaysbelessthanorequaltoPynet.

Agroupof18columnspredisposedtoaglobalfailurewereselectedfromtheSSMA

columnsimulationdatabase.Inthisstudythecolumnlength,L,variedfrom8in.to96
in. to consider a wide range of global column slenderness, c=(Pyg/Pcre)0.5. The SSMA
crosssections are purposely selected with low local buckling slenderness (i.e., sections
withthickersheetsteeluptot=0.1017in.andrelativelynarrowflangesandwebs).DSM
predictsthatlocalbucklingdoesnotinfluenceglobalbucklingbehaviorwhen l0.776.
ThewebofeachcolumncontainsevenlyspacedslottedholeswheretheholespacingS
variesfrom8in.to22in.Theholelength,Lhole,isheldconstantat4in.,whilethehole
depth, hhole, is varied for each column to produce Anet/Ag of 1.0 (no holes), 0.9, and 0.8.
(Refer to Appendix K, Study Type G, for specific column crosssection and hole
1695H

geometry information.) The four columns with the lowest global slenderness (for
example, Specimen ID # 137 to 140 in Appendix K) were modeled with circular holes
169H

insteadofslottedholesbecausetheslottedholesresultedinimpracticalcolumnlayouts,

323

with the hole extending over more than 50% of the column length. The global
imperfectionshapeforfiveofthelongercolumnswasweakaxisflexuralbuckling,and
therefore four simulated strengths are determined for these columns (instead of the
typical two): 25% CDF local and distortional imperfections with L/2000 global
imperfections and 75% CDF local and distortional imperfections with L/1000 global
imperfections.
Figure8.9to Figure8.11comparethesimulatedcolumnstrengthstotheDSMglobal

1697H

1698H

predictioncurveasAnet/Agdecreases.Thesimulatedstrengthsforcolumnswithoutholes
areconsistentwiththeDSMglobalpredictioncurveasshownin Figure8.9a.Themean
169H

and standard deviation of the simulated test to predicted ratio for columns without
holes is 1.06 and 0.05 respectively for 25% CDF local and distortional imperfections
L/2000 global imperfection and 0.95 and 0.07 for 75% local and distortional
imperfectionsL/1000globalimperfection.

Figure8.10aandFigure8.11ademonstratethatforcolumnswithholes,thepredicted
170H

170H

strengthsareconsistentwiththeDSMglobaldesigncurvewhenglobalslenderness cis
greater than 2. Most of the columns in this region fail by weakaxis flexural buckling.
Whencisbetween1and2,allofthecolumnsfailbyflexuraltorsionalbucklingandthe
simulatedcolumnstrengths(with25%CDFimperfections)are20%higherthantheDSM
predictions. This conservative trend is caused by the simplified prediction method
developed in Section 4.3.2.3, which is know tobe a conservative predictor of Pcrewhen
1702H

torsionalbucklinginfluencestheglobalbucklingmode.WhenPcreisunderpredicted,the
global slenderness increases, which shifts the tested data off of the DSM design curve;

324

theshiftisespeciallyclearinFigure8.11a.Thisobservationfurthermotivatesthefuture
1703H

worktostudytheinfluenceofholesonthewarpingtorsionconstant,Cw.
Simulated column strengths diverge below the DSM prediction curve when c

decreases and Anet/Ag increases as shown in Figure 8.10a and Figure 8.11a. These
1704H

1705H

columns are short, ranging in length from 8 in.to 26 in., and exhibit a yielding failure
mode at the net section, similar to that observed in the distortional failure study in
Figure 8.3. This observation supports the proposed modification to the DSM
1706H

distortionalbucklingcurve,whichaccuratelypredictsthestrengthsofthesecolumnsas
shown in Figure 8.9b and Figure 8.11b, where the diverging data points are plotted
170H

1708H

againstthemodifiedDSMdistortionalpredictioncurve.Thisobservationreiteratesthe
conclusion drawn in the distortional buckling study, that yielding and collapse of the
unstiffenedstripsadjacenttoaholeinfluencebothdistortionalandglobalfailuremodes
asslendernessdecreases.
1.4

1.4
DSM (no hole)
no holes, FE 25% CDF imperfections
no holes, FE 75% CDF imperfections

0.8

0.8

0.6

0.6

0.4

0.4

0.2

0.2

0.5

1.5

2.5

3.5

DSM (no hole)


no holes, FE 25% CDF imperfections
no holes, FE 75% CDF imperfections

1.2

Ptest/Py

Ptest/Py

1.2

global slenderness, c =(Py /Pcre)0.5

0.5

1.5

2.5

distortional slenderness, d=(Py /P crd)0.5

3.5

Figure8.9Comparisonofsimulatedcolumnstrengths(Anet/Ag=1.00)to(a)theexistingDSMglobalbuckling
designcurveandto(b)theproposedDSMdistortionalbucklingcurveforcolumnswithholes

325

1.4

1.4
DSM (no hole)
slotted holes, FE 25% CDF imperfections
slotted holes, FE 75% CDF imperfections
circular holes, FE 25% CDF imperfections
circular holes, FE 75% CDF imperfections

1.2

0.8

Ptest/Py

Ptest/Py

0.6

0.8

0.6

0.4

0.4

0.2

0.2

0.5

1.5

2.5

3.5

DSM (no hole)


slotted holes, FE 25% CDF imperfections
slotted holes, FE 75% CDF imperfections

1.2

0.5

1.5

2.5

3.5

distortional slenderness, d=(Py /P crd)0.5

global slenderness, c =(Py /Pcre)0.5

Figure8.10Comparisonofsimulatedcolumnstrengths(Anet/Ag=0.90)to(a)theexistingDSMglobalbuckling
designcurveandto(b)theproposedDSMdistortionalbucklingcurveforcolumnswithholes
1.4

1.4
DSM (no hole)
slotted holes, FE 25% CDF imperfections
slotted holes, FE 75% CDF imperfections
circular holes, FE 25% CDF imperfections
circular holes, FE 75% CDF imperfections

1.2

0.8

Ptest/Py

Ptest/Py

0.6

0.8

0.6

0.4

0.4

0.2

0.2

0.5

1.5

2.5

3.5

DSM (no hole)


slotted holes, FE 25% CDF imperfections
slotted holes, FE 75% CDF imperfections

1.2

0.5

1.5

2.5

distortional slenderness, d=(Py /P crd)0.5

global slenderness, c =(Py /Pcre)0.5

3.5

Figure8.11Comparisonofsimulatedcolumnstrengths(Anet/Ag=0.80)to(a)theexistingDSMglobal
bucklingdesigncurveandto(b)theproposedDSMdistortionalbucklingcurveforcolumnswithholes

1.4 Local buckling study


106B

The distortional buckling failure study in Section 8.1.2 and the global buckling
1709H

failurestudyinSection 8.1.3demonstratedthatthepresenceofholesdecreasesultimate
170H

strengthwhencoldformedsteelcolumnsfailbyyieldingandcollapseoftheunstiffened
stripsadjacenttoaholeatthenetcrosssection.Holeswereobservedtohaveaminimal
influence on ultimate strength when the column failure mode was dictated by elastic
buckling. Thegoalofthisstudyistodetermineifthistrendisconsistentforcolumns
withholesexperiencinglocalglobalbucklinginteractionatfailure.

326

Elevencolumnsfromthesimulationdatabasein AppendixKwerechosenforthis
17H

study. The columns have SSMA crosssections and lengths which result in a local
bucklingslenderness, l,rangingfrom0.8to3.0.Thecolumnlength,L,variesfrom24
in. to 88 in. and column widths range from 3.5 in. to 12 in. The web of each column
containsevenlyspacedcircularholeswheretheholespacingSvariesfrom12in.to17
in.Theholedepth(diameter),hhole,isvariedforeachcolumntoproduceAnet/Agof1.0(no
holes),0.80,and0.65.Referto AppendixK,StudyTypeL,forspecificcolumncross
172H

sectionandholegeometryinformation.

Thesimulatedultimatestrengthsofthe11columnswithoutholes,Ptest,arecompared

totheDSMlocalbucklingstrengthprediction,Pnl,in Figure8.12.Thesimulatedtestto
173H

predicted ratios are more variable than those observed in the distortional and global
failure studies but on average are close to unity, with a trend towards increasingly
conservative predictions with increasing l as shown in Figure 8.12a. The mean and
174H

standarddeviationofthesimulatedtesttopredictionratiois1.05and0.14respectively
for25%CDFlocalanddistortionalimperfectionsL/2000globalimperfectionsand1.03
and0.15for75%localanddistortionalimperfectionsL/1000globalimperfections.

327

1.4

1.2

1.2

0.8

Ptest/Pn

Ptest/Pn

1.4

no holes, FE 25% CDF imperfections


no holes, FE 75% CDF imperfections

0.6

0.8

0.4

0.4

0.2

0.2

0.5

1.5

2.5

3.5

no holes, FE 25% CDF imperfections


no holes, FE 75% CDF imperfections

0.6

local slenderness, =(Pne/Pcr )0.5


l
l

0.5

1.5

2.5

3.5

global slenderness, c =(Py /Pcre)0.5

Figure8.12Comparisonofcolumntesttopredictionratiosforcolumns(Anet/Ag=1.0)failingbylocalglobal
bucklinginteractionasafunctionof(a)localslenderness(b)globalslenderness

Figure8.13andFigure8.14comparethesimulatedstrengthsofthe11columnstothe
175H

176H

predictedstrength,Pnl,asAnet/Agdecreasesfrom1.0(nohole),to0.80,to0.65.(InFigure
17H

8.13andFigure8.14theteststrengthsarethoseassociatedwithPtest25+inAppendixK,i.e.
178H

179H

the 25% CDF local and distortional imperfection magnitudes and +L/2000 global
imperfection magnitudes.) Figure 8.13a compares the simulated strengths Ptest25+ to Pnl
1720H

without the influence of holes. (The localglobal buckling interaction complicates the
comparisonbecausePnlandlarebothafunctionofPne.ByinitiallyassumingthatPneis
notinfluencedbythehole,theeffectofholesizeonsimulatedstrengthismoreclearly
observed.)Aslocalslenderness(l)decreasesin Figure8.13a(i.e.,theinfluenceoflocal
172H

bucklingonmemberstrengthdecreases)thetestedstrengthbecomesmoresensitiveto
increasing hole size (i.e., decreasing Anet/Ag), diverging below the prediction Pnl by as
muchas40%whenl=0.75.
Figure8.14ademonstratesthatthesensitivityofcolumnstrengthtoadecreaseinAnet/Ag
172H

isrelatedtotheratioofPynettoPne.WhenPynet/Pneishigh,thestrengthsensitivitytoAnet/Ag

328

islowbecauseglobalbucklinginitiatesthecolumnfailure.AsPne/Pynetapproachesunity,
column failure is initiated by unstiffened strip buckling and yielding at the net cross
section and therefore the sensitivity of column strength to Anet/Ag increases. A column
with the largest drop in strength with increasing hole size is the SSMA 350S16268
column with L=34 in. and S=17 in. shown in Figure 8.15. In this case a large hole
1723H

(Anet/Ag=0.65) causes the collapse of the net section resulting in an unfavorable and
suddenweakaxisflexuralfailureanda42%strengthreductionwhencomparedtothe
same column without holes. The SSMA 350S16268 column with smaller holes
(Anet/Ag=0.80)failsinacombinationofdistortionalandflexuraltorsionalbucklingwitha
12%strengthreduction.

Figure8.13bplotsthesameinformationas Figure8.14a,exceptnowPneiscalculated
1724H

1725H

usingPcre includingtheinfluenceofholes.For8outofthe11columns,theprediction
Pnl shifts from unconservative to slightly conservative, even for large holes. One
exception is the SSMA 800S25043 column with L=74 in. and S=12 in. shown in Figure
1726H

8.16,wherethestrengthpredictionbecomesoverlyconservativeasAnet/Agincreases.Pcre
ispredictedtodecreaseby45%whenAnet/Ag=0.65,althoughthetestedstrengthdecreases
by only 10%. Figure 8.16 demonstrates that the Csection web is susceptible to local
172H

buckling, and that the presence of holes does not adversely affect the failure mode in
this case. The strengthpredictions for the SSMA 350S16268 column (Figure 8.15)and
1728H

theSSMA350S16254columnwithL=24in.andS=12in.areviablewhenAnet/Ag=0.80,but
are underestimated by 20% with Option 4 (5) when Anet/Ag=0.65 because of the
introductionofanunstableweakaxisflexuralfailuremodetriggeredbythecollapseof

329

the net section. A hinge forms at the location of the net section, and the global
slendernessishighenoughthatthecolumnbecomessusceptibletoaflexural buckling
mode. These hinge failures are not observed in the distortional buckling study (see
Section 8.1.2)becausetheglobalslendernessofthecolumnsislower(i.e.theweakaxis
1729H

flexural stiffness is higher), avoiding a global buckling failure. Option 6 accurately


predicts the strength of the SSMA 350S16268 column and the SSMA 350S16254
columnsbecausethemethodassumesthattheglobalstrength,Pne,isreducedbyPynet/Py.

1.5

SSMA 800S250-43 column

1.5

Anet/Ag
1.0
1

1
Ptest/Pn

Ptest/Pn

0.80
0.65

0.5

0.5

SSMA 350S162-68 column

0.5

1.5

2.5

3.5

local slenderness, =(Pne/Pcr )0.5


l
l

0.5

1.5

2.5

3.5

local slenderness, =(P ne/Pcr )0.5


l
l

Figure8.13Comparisonofcolumntesttopredictionratiosforcolumnsfailingbylocalglobalbuckling
interactionwithPnecalculated(a)withouttheinfluenceofholes(b)andwiththeinfluenceofholes

330

1.5

1
Ptest/Pn

P test/Pn

1.5

Increase in point size,


Anet/Ag=1.0,0.80,0.65

0.5

0.5

1.5

2.5
P ynet/Pne

3.5

0.5

4.5

0.5

1.5

2.5
P ynet/Pne

3.5

4.5

Figure8.14Comparisonofcolumntesttopredictionratiosforcolumnsfailingbylocalglobalbuckling
interactionasafunctionofPynet/PnewherePneiscalculated(a)withouttheinfluenceofholes(b)andwiththe
influenceofholes

Distortional,
flexural-torsional failure

Distortional,
flexural-torsional failure

Weak-axis flexural failure

Anet/Ag

1.0

0.80

0.65

Ptest/Ptest,no hole

1.0

0.88

0.58

Figure8.15SSMA350S16268columnfailuremodechangesfromdistortionalflexuraltorsionalbuckling
failuretoweakaxisflexureasholesizeincreases(L=34in.)

331

Local global interaction


at failure

Local global interaction


at failure

Unstiffened strip buckling


global interaction at failure

Anet/Ag

1.0

0.80

0.65

Ptest/Ptest,no hole

1.0

0.95

0.90

Figure8.16SSMA800S25043(L=74in.)columnweblocalbucklingchangestounstiffenedstripbucklingat
peakloadasholesizeincreases

Theobservationsfromthisstudyarenowemployedtoproposetwooptionsforthe

DSM local buckling design curve for columns with holes. The presence of holes
influenced the tested strength of the coldformed steel columns over the full range of
localslendernessconsidered.Thisresultwasdifferentfromthedistortionalandglobal
failurestudies,whereholeswereobservedtoreducestrengthonlyfromthecollapseand
yielding at the net section as slenderness decreased. The strength reduction from the
holeswaspredictedinDSMfor8outofthe11columns,whenPnlwascalculatedwith
Pneincludedtheinfluenceofholes(compare Figure8.13ato Figure8.13b).Atransition
1730H

173H

similartothatproposedfortheDSMdistortionaldesigncurveisstill justifiedthough,
especiallywhenPynet/Pne1(see Figure8.14b),tocapturetheyieldingandcollapseatthe
1732H

netsectionobservedincolumnswithlowlocalandglobalslenderness.Thestrengthof
twocolumnswithlargeholeswereunderpredictedbecauseofunstableglobalcollapse
initiatedbyyieldingatthenetsection,motivatingtheimplementationofalimitonhole

332

size(i.e.,Anet/Ag)toensuretheviabilityoftheDSMapproach.Twomodificationoptions
areproposedfortheDSMlocaldesigncurvebasedontheseconclusions:
Local Buckling (Option A)
The nominal axial strength, Pnl, for local buckling shall be calculated in accordance with the following:
(a) For l l1

Pnl = Pne Pynet

(b) For l1<ll2

(caponcolumnstrength)

P
( l l1 )

ynet
l2
Pnl = P

ynet

l1
l2
P

(c) For l > l2

Pnl = 1 0.15 Pcrl

0.4

Pne

where
l

P
crl
Pne

(yieldtransitionwhenPynet/Pne1)

0. 4

Pne

(DSMlocalbucklingcurve,unchanged)

= Pne Pcrl

l1

= 0.776(Pynet Pne ) 0.776

l2 = 0.776 1.7 (Pynet Pne )1.6 0.7 , Pynet/Pne1

= 0.776, Pynet/Pne>1
= 1 0.15(1 l 2 )0.8 (1 l 2 )0.8 Pne
Pynet = FyAnet0.6Py
Pl2

(notransitionwhenPynet/Pne>1)
(limitreductionofthenetsectionto0.6Py)

Anet = Column cross-sectional area at the location of hole(s)


Pcrl = Critical elastic local column buckling load including hole(s)

Local Buckling (Option B)


The nominal axial strength, Pnl, for local buckling shall be calculated in accordance with the following:
(a) For l l1

Pnl = Pynet (Pne /Py )

(caponcolumnstrength)

(b) For l1<ll2

Pynet (Pne Py ) Pl 2
(l l1 )
P
l 2 l1

Pnl = P Pne
ynet

(c) For l > l2

Pnl = 1 0.15 Pcrl

Pne

where
l
l1

l2

0.4

P
crl
Pne

0. 4

(DSMlocalbucklingcurve,unchanged)

Pne

= Pne Pcrl
= 0.776(Pynet Py )

= 0.776 1.7 (Pynet Py )1.6 0.7

= 1 0.15(1 l 2 )
Pynet = FyAnet0.6Py
Pl2

(yieldtransitionwhenPynet/Pne1)

0.8

)(1

l2

0.8

Pne
(limitreductionofthenetsectionto0.6Py)

333

Anet = Column cross-sectional area at the location of hole(s)


Pcrl = Critical elastic local column buckling load including hole(s)

Option A imposes a transition from the DSM local buckling curve to column

strengthatthenetsection,Pynet,whenPynet<PneasshowninFigure8.17aforthecasewhen
173H

Pne=Py (i.e., stub columns) and Pynet=0.8Pyg. When Pynet>Pne, Option A assumes that holes
influence only the critical elastic buckling loads (Pcrl, Pcre) and otherwise do not change
thefailuremodeofthecolumn;thiscaseisdemonstratedinFigure8.17cwhenPcre=Pyg.
1734H

OptionBalsoimposesatransitiontothenetcolumnstrengthfromtheDSMlocalfailure
curve,althoughthetransitionisassumedtooccurforall valuesofPynet/Pne.Inessence,
theOptionBcurveforstubcolumnsshownin Figure8.17aisscaleddownbasedonthe
1735H

ratio Pynet/Py. The result is an additional reduction in predicted strength for global
column failures without local buckling interaction that is not captured by Option A.
This difference between Option A and Option B is highlighted in Figure 8.17b, where
1736H

Pynet=0.8Pyg and Pcre = 5Pyg. The validity of both options are evaluated in the following
section against the simulation database and the experiment database assembled in
Chapter4.
173H

1.4

1.4
DSM local curve (no hole)
DSM local curve (Option A)
DSM local curve (Option B)

1.2

1.4
DSM local curve (no hole)
DSM local curve (Option A)
DSM local curve (Option B)

1.2

Pcre=100Pyg

Pcre=5Pyg
0.8

0.6

0.6

0.4

0.4

0.4

0.2

0.2

0.2

0.5

1.5

2.5

local slenderness, =(P /P )0.5


ne cr l
l

3.5

0.5

1.5

2.5

local slenderness, =(P /P )0.5


ne cr l
l

Pynet=0.8Pyg

Pn /Py

0.6

Pcre=Pyg
0.8

Pynet=0.8Pyg

Pn /Py

Pynet=0.8Pyg

Pn /Py

0.8

DSM local curve (no hole)


DSM local curve (Option A)
DSM local curve (Option B)

1.2

3.5

0.5

1.5

2.5

3.5

local slenderness, =(P /P )0.5


ne cr l
l

Figure8.17ComparisonofDSMlocalbucklingdesigncurveoptionswhenPynet=0.80Pygand(a)Pcre=100Pyg,
(b)Pcre=5Pyg,and(c)Pcre=Pyg

334

1.5 Presentation and evaluation of DSM options


107B

SixoptionsforextendingDSMtocolumnswithholesareevaluatedinthissection.

The options range from simple substitutions in the existing code to more involved
modifications, including the incorporation of the design curve transitions discussed in
Section8.1.2andSection8.1.4fordistortionalandlocalbuckling.
1738H

1739H

1.5.1 DescriptionofDSMoptions
182B

Option 1: Include hole(s) in Pcr determinations, ignore hole otherwise


This method, in presentation, appears identical to currently available DSM expressions
Flexural, Torsional, or Torsional-Flexural Buckling
The nominal axial strength, Pne, for flexural, or torsional- flexural buckling is
2
for c 1.5 Pne = 0.658 c Py

0.877
for c > 1.5 Pne = 2 Py = 0.877Pcre

c
where

Py

= AgFy

Py Pcre

Pcre= Critical elastic global column buckling load (including hole(s))


Ag

= gross area of the column

Local Buckling
The nominal axial strength, Pnl, for local buckling is
for l 0.776 Pnl = Pne

335


P
for l > 0.776 Pnl = 1 0.15 crl

Pne

where

0 .4

P
crl
Pne

0.4

Pne

Pne Pcrl

Pcrl = Critical elastic local column buckling load (including hole(s))


Pne is defined above.
Distortional Buckling
The nominal axial strength, Pnd, for distortional buckling is
for d 0.561 Pnd = Py

for d > 0.561 Pnd = 1 0.25 crd


Py

where

0 .6

Pcrd
P
y

0 .6

Py

= Py Pcrd

Pcrd = Critical elastic distortional column buckling load (including hole(s))

336

Option 2: Include hole(s) in Pcr determinations, Use Pynet everywhere


The only change in this method is to replace Py with Pynet
Flexural, Torsional, or Torsional-Flexural Buckling
The nominal axial strength, Pne, for flexural, or torsional- flexural buckling is
for cnet 1.5 Pne = (0.658
for cnet > 1.5
where

cnet

2
cnet

)P

ynet

0.877
Pne = 2 Pynet = 0.877 Pcre
cnet

Pynet Pcre

Pynet = AnetFy
Pcre= Critical elastic global column buckling load (including hole(s))
= net area of the column

Anet

Local Buckling
The nominal axial strength, Pnl, for local buckling is
for l 0.776 Pnl = Pne

P
for l > 0.776 Pnl = 1 0.15 crl

Pne

where

0 .4

P
crl
Pne

0.4

Pne

Pne Pcrl

Pcrl = Critical elastic local column buckling load (including hole(s))


Pne is defined above
Distortional Buckling
The nominal axial strength, Pnd, for distortional buckling is
for dnet 0.561 Pnd = Pynet

for dnet > 0.561 Pnd = 1 0.25 crd


Pynet

where

dnet

0.6

Pcrd

Pynet

0.6

Pynet

= Pynet Pcrd

Pcrd = Critical elastic distortional column buckling load (including hole(s))

337

Option 3: Cap Pnl and Pnd, otherwise no strength change, include hole(s) in Pcr
This method puts bounds in place and assumes local-global interaction happens at full Pne
Flexural, Torsional, or Torsional-Flexural Buckling
The nominal axial strength, Pne, for flexural, or torsional- flexural buckling is
2
for c 1.5 Pne = 0.658 c Py

0.877
for c > 1.5 Pne = 2 Py = 0.877Pcre

c
where

Py Pcre

Pcre= Critical elastic global column buckling load (including hole(s))


Py
Ag

= AgFy
= gross area of the column

Local Buckling
The nominal axial strength, Pnl, for local buckling is
for l 0.776 Pnl = Pne Pynet

P
for l > 0.776 Pnl = 1 0.15 crl

Pne

where

0 .4

Pcrl
Pne

0 .4

Pne Pynet

Pne Pcrl

Pcrl = Critical elastic local column buckling load (including hole(s))


Pne is defined in Section above.
Pynet = AnetFy
Anet

= net area of the column

Distortional Buckling
The nominal axial strength, Pnd, for distortional buckling is
for d 0.561 Pnd = Py Pynet

for d > 0.561 Pnd = 1 0.25 crd


Py

where

0.6

Pcrd

Py

0.6

Py P.ynet

= Py Pcrd

Pcrd = Critical elastic distortional column buckling load (including hole(s))

338

Option 4: Cap Pnl, transition Pnd, include hole(s) in Pcr determinations


This method puts bounds and transition in place, assumes local-global interaction at full Pne
Flexural, Torsional, or Torsional-Flexural Buckling
The nominal axial strength, Pne, for flexural, or torsional- flexural buckling is
2
for c 1.5 Pne = 0.658 c Py

0.877
for c > 1.5 Pne = 2 Py = 0.877Pcre

c
where

Py Pcre

Pcre= Critical elastic global column buckling load (including hole(s))


Py
Ag

= AgFy
= gross area of the column

Local Buckling
The nominal axial strength, Pnl, for local buckling is
for l 0.776 Pnl = Pne Pynet

P
for l > 0.776 Pnl = 1 0.15 crl

Pne

where

0 .4

Pcrl
Pne

0 .4

Pne Pynet

Pne Pcrl

Pcrl = Critical elastic local column buckling load (including hole(s))


Pne is defined in Section above.
Pynet = AnetFy
Anet = net area of the column
Distortional Buckling
The nominal axial strength, Pnd, for distortional buckling shall be calculated in accordance with the
following:
(a) For d d1
Pnd = Pynet

(b) For d1<dd2


Pnd = P Pynet Pd 2
ynet

d1
d2

(c) For d > d2

Pnd = 1 0.25 crd


Py

where

0 .6

Pcrd
P
y

0 .6

d1

= 0.561(Pynet Py )

Py

Py Pcrd

d2

= 0.561 14 (Pynet Py )0.4 13

Pd2

= 1 0.25 (1 d 2 )1.2 (1 d 2 )1.2 Py

Pynet = FyAnet0.6Py
Anet = Column cross-sectional area at the location of hole(s)
Pcrd

= Critical elastic distortional column buckling load including hole(s)

339

Option 5: Transition Pnl (Option A), transition Pnd, include hole(s) in Pcr determinations
This method puts bounds and transition in place, assumes local-global interaction at full Pne
Flexural, Torsional, or Torsional-Flexural Buckling
The nominal axial strength, Pne, for flexural, or torsional- flexural buckling is
2
for c 1.5 Pne = 0.658 c Py

0.877
for c > 1.5 Pne = 2 Py = 0.877Pcre

c
where

Py Pcre

Pcre= Critical elastic global column buckling load (including hole(s))


Py

= AgFy
= gross area of the column

Ag

Local Buckling
The nominal axial strength, Pnl, for local buckling shall be calculated in accordance with the following:
(a) For l l1

Pnl = Pne Pynet

(b) For l1<ll2

(caponcolumnstrength)

P
( l l1 )

ynet
l2
Pnl = P

ynet
l 2 l1
P

(c) For l > l2

Pnl = 1 0.15 Pcrl

0.4

Pne

where
l

l1

P
crl
Pne

(yieldtransitionwhenPynet/Pne1)

0. 4

(DSMlocalbucklingcurve,unchanged)

Pne

= Pne Pcrl
= 0.776(Pynet Pne ) 0.776

l2 = 0.776 1.7 (Pynet Pne )1.6 0.7 , Pynet/Pne1

= 0.776, Pynet/Pne>1
= 1 0.15(1 l 2 )0.8 (1 l 2 )0.8 Pne
Pynet = FyAnet0.6Py
Pl2

(notransitionwhenPynet/Pne>1)
(limitreductionofthenetsectionto0.6Py)

Anet = Column cross-sectional area at the location of hole(s)


Pcrl = Critical elastic local column buckling load including hole(s)
Distortional Buckling

SameasOption4

340

Option 6: Transition Pnl (Option B), transition Pnd, include hole(s) in Pcr determinations
This method puts bounds and transition in place, assumes local-global interaction at full Pne
Flexural, Torsional, or Torsional-Flexural Buckling
The nominal axial strength, Pne, for flexural, or torsional- flexural buckling is
2
for c 1.5 Pne = 0.658 c Py

0.877
for c > 1.5 Pne = 2 Py = 0.877Pcre

c
where

Py Pcre

Pcre= Critical elastic global column buckling load (including hole(s))


Py

= AgFy
= gross area of the column

Ag

Local Buckling
The nominal axial strength, Pnl, for local buckling shall be calculated in accordance with the following:
(a) For l l1

Pnl = Pynet (Pne /Py )

(caponcolumnstrength)

(b) For l1<ll2

Pynet (Pne Py ) Pl 2
(l l1 )
P
l 2 l1

Pnl = P Pne
ynet

(c) For l > l2

Pnl = 1 0.15 Pcrl

Pne

where
l

l1
l2

0.4

Pl2

P
crl
Pne

(yieldtransitionwhenPynet/Pne1)

0. 4

Pne

(DSMlocalbucklingcurve,unchanged)

= Pne Pcrl
= 0.776(Pynet Py )

= 0.776 1.7 (Pynet Py )1.6 0.7

= 1 0.15(1 l 2 )

Pynet = FyAnet0.6Py

0.8

)(1

l2

0.8

Pne
(limitreductionofthenetsectionto0.6Py)

Anet = Column cross-sectional area at the location of hole(s)


Pcrl = Critical elastic local column buckling load including hole(s)
Distortional Buckling

SameasOption4

341

1.6 DSM comparison to column test simulation database


108B

The six DSM prediction options for coldformed steel columns with holes are

evaluated with the simulated column experiment database developed in Section 8.1.1
1740H

and summarized in Appendix K. (Tested strengths with and without global


174H

imperfectionsareprovidedin AppendixK.Thesimulatedstrengthsconsideredinthis
1742H

study contain global imperfections, except for stocky columns with L/D<18whereD is
the column flange width). The simulated data is compared against DSM predictions
whileevaluatingdatatrendsagainstmemberslenderness,holesize(Anet/Ag),andcolumn
dimensionsL/h,wherehistheflatwebwidthofacolumn.

Figure 8.18 to Figure 8.21 compare the simulated test data to predictions for local,
1743H

174H

distortional,andglobalbucklingcontrolledcolumnfailures.Option1isidenticaltothe
existing DSM approach for columns without holes, except the critical elastic buckling
loads(Pcrl,Pcrd,andPcre)aredeterminedwiththeinfluenceofholes.Option1isobserved
to be an accurate predictor of strength when l, d, and c are high, but results in
unconservative predictions (by as much as 30 % for distortional buckling controlled
specimens, see Figure 8.20) as l, d, and c decreases below 1.5. The unconservative
1745H

predictionsoccurbecauseOption1doesnotaccountforthecolumnstrengthlimitPynet,
nordoesitaccountforatransitionfromanelasticbucklingcontrolledfailuretoayield
controlledfailureatthenetsectiondiscussedinSection8.1.2andSection8.1.3.
1746H

174H

Option2isobservedtobeaconservativepredictorin Figure8.18to Figure8.21for


1748H

1749H

highl,d,andcanddemonstratesimprovedaccuracyoverOption1whenslenderness
decreases and hole size increases (see Figure 8.20). Option 2 replaces Pynet everywhere
1750H

342

withintheexistingDSMformulation,whichhastheeffectofincreasingl,d,andcand
decreasingpredictedstrength.Option3testtopredictedtrendsaresimilartoOption1
with increasingly unconservative predictions as slenderness decreases, demonstrating
thatthePynetlimitonPnlandPndinOption3arenotfullyeffectiveatcapturingtheyield
transitiontothenetsection.Option4isidenticaltoOption3excepttheyieldtransition
ontheDSMdistortionalcurvedevelopedinSection8.1.2isemployedtoprovideamore
175H

accurate prediction of the netsection yielding influence. Option 4 demonstrates an


improvementinaccuracyoverOption3,althoughitoverpredictsthestrengthofthetwo
columnsdiscussedinSection 8.1.4(SSMA350S16268andSSMA350S16254columns),
1752H

where large holes caused a sudden weakaxis flexural buckling failure. Option 5
includes both local and distortional yield transitions, although the predictions are
identicaltoOption4becausethedistortionaltransitionalwayspredictslowerstrengths
thanthelocaltransitionforthecolumnsconsidered.Option6deviatesfromtheother
approaches and accounts for the presence of holes by reducing Pnl by the ratio Pynet/Py
when lislessthan0.776;thisoptionalsoalwaysincludingalocalbucklingtransition
(Option 5 imposes a transition on the DSM local buckling design curve only when
Pynet<Pne, see Figure 8.17). The reduction in Pnl shifts the global bucklingcontrolled
1753H

specimensinOptions1through5totheDSMlocalbucklingcurveinOption6,resulting
inconservativepredictionswithdecreasingl.

Table 8.1 summarizes the testtopredicted ratio statistics for the six DSM options.
1754H

Thestandarddeviation(SD)isusefulwhencomparingthemethods,becauseitprovides
a metric for how well the trends in strength are following the prediction curves. (The

343

mean is also an important statistic but can hide unconservative prediction trends in
some columns with overconservative predictions in other columns). A low standard
deviation is appealing because it enables higher strength reduction factors in a design
code.Thestrengthreductionfactor isalsoprovidedforeachoption. iscalculated
withthefollowingequationfromChapterFoftheSpecification(AISIS1002007):

= C (M m Fm Pm )e

o VM2 +VF2 + C PVP2 +VQ2

(8.1)

wherethecalibrationcoefficientC =1.52forLRFD,themeanvalueofthematerialfactor
Mm=1.10 for concentrically loaded compression members, the mean value of the
fabricationfactorFm=1.0,themeanvalueoftheprofessionalfactorPm=1.0,thecoefficient
of variation (COV) of the material factor Vm=0.10, the COV of the fabrication factor
Vf=0.05,theCOVoftheloadeffectVq=0.21forLRFD,andthecorrectionfactorCp=1.The
COV of the test results, Vp, is calculated as the ratio of the standard deviation to the
meanofthetesttopredictedstatisticsinTable8.1.
175H

Nooneoptionstandsoutabovetherestwhenstudyingthetable,althoughOption2,

3, and 4 (5) have the most evenly distributed statistics between local and distortional
buckingcolumngroups.Theobservationsfromthiscomparisonwillbecombinedwith
theDSMcomparisontotheexperimentaldatabaseinthenextsection.
Table8.1DSMtesttopredictedstatisticsforcolumnsimulations

Option
1

Description
Py everywhere

Mean
1.06

Local buckling
SD

0.15
0.83

# of tests
93

Mean
1.07

Distortional buckling
SD
# of tests

0.17
0.82
178

Mean
1.11

Global buckling
SD

0.21
0.78

# of tests
114

Pynet everywhere

1.14

0.13

0.86

93

1.24

0.18

0.83

176

1.15

0.18

0.82

116

Cap Pnl, Pnd

1.06

0.15

0.83

93

1.09

0.17

0.82

186

1.13

0.21

0.79

106

Transition Pnd, Cap Pnl

1.08

0.14

0.85

89

1.04

0.19

0.79

200

1.16

0.19

0.82

96

Transition Pnd and Pnl (Option A)

1.08

0.14

0.85

89

1.04

0.19

0.79

200

1.16

0.19

0.82

96

Transition Pnd and Pnl (Option B)

1.07

0.20

0.78

221

1.10

0.15

0.85

164

---

---

---

344

1.5

1.5
Option 1 - Py everywhere

Option 2 - Pynet everywhere


1
Pn/Pne

Pn/Pne

0.5

0.5

1.5

2.5

0.5

0.5

0.5

Option 4 - cap Pn , transition Pnd

1
Pn/Pne

Pn/Pne

DSM Local Prediction


Local Controlled Test

0.5

0.5

1.5

2.5

0.5

0.5

0.5

1.5

2.5

2.5

0.5

l=(Pne/Pcrl)

l=(Pne/Pcrl)

1.5

1.5
Option 5 - transition Pn (Option A), transition Pnd
l

Option 6 - transition Pn (Option B), transition Pnd


l
1
Pn/Pne

1
Pn/Pne

2.5

1.5
Option 3 - cap Pn , Pnd

0.5

l=(Pne/Pcrl)

1.5

1.5
0.5

l=(Pne/Pcrl)

0.5

1.5

2.5

0.5

0.5

0.5

1.5

2
0.5

l=(Pne/Pcrl)

l=(Pne/Pcrl)

Figure8.18Comparisonofsimulatedteststrengthstopredictionsforcolumnswithlocalbucklingcontrolledfailuresasafunctionoflocalslenderness(tested
strengthisnormalizedbyPne)

345

1.5

1.5
Option 1 - Py everywhere

Option 2 - Pynet everywhere

Pn/Py

Pn/Py

0.5

0.5

0.5

1.5

2.5

0.5

0.5

1.5

2.5

0.5

l=(Pne/Pcrl)

l=(Pne/Pcrl)

1.5

1.5
Option 3 - cap Pn , Pnd

Option 4 - cap Pn , transition Pnd

DSM Local Prediction


Local Controlled Test

Pn/Py

Pn/Py

0.5

0.5

0.5

1.5

2.5

0.5

0.5

1.5

2.5

2.5

0.5

l=(Pne/Pcrl)

l=(Pne/Pcrl)

1.5

1.5
Option 5 - transition Pn (Option A), transition Pnd
l

Option 6 - transition Pn (Option B), transition Pnd


l
Pn/Py

Pn/Py

0.5

0.5

0.5

1.5

2.5

0.5

0.5

1.5

2
0.5

l=(Pne/Pcrl)

l=(Pne/Pcrl)

Figure8.19Comparisonofsimulatedteststrengthstopredictionsforcolumnswithlocalbucklingcontrolledfailuresasafunctionoflocalslenderness(tested
strengthisnormalizedbyPyg)

346

1.5

1.5
Option 1 - Py everywhere

Option 2 - Pynet everywhere

Pn/Py

Pn/Py

DSM Dist. Prediction


Dist. Controlled Test

0.5

0.5

0.5

1.5

2.5

0.5

0.5

1.5

2.5

2.5

2.5

0.5

d=(Py /Pcrd)

dnet=(Pynet/Pcrd)

1.5

1.5
Option 3 - cap Pn , Pnd

Option 4 - cap Pn , transition Pnd

Pn/Py

Pn/Py

0.5

0.5

0.5

1.5

2.5

0.5

0.5

1.5
0.5

d=(Py /Pcrd)

d=(Py /Pcrd)

1.5

1.5
Option 5 - transition Pn (Option A), transition Pnd
l

Option 6 - transition Pn (Option B), transition Pnd


l
Pn/Py

Pn/Py

0.5

0.5

0.5

1.5

2.5

0.5

0.5

1.5

2
0.5

d=(Py /Pcrd)

d=(Py /Pcrd)

Figure8.20Comparisonofsimulatedteststrengthstopredictionsforcolumnswithdistortionalbucklingcontrolledfailuresasafunctionofdistortional
slenderness

347

1.5

1.5
Option 2 - Pynet everywhere

Option 1 - Py everywhere

Pn/Py

Pn/Py

DSM Global Prediction


Global Controlled Test

0.5

0.5

0.5

1.5

2.5

0.5

0.5

1.5

2.5

2.5

2.5

0.5

c =(Py /Pcre)

cnet=(Pynet/Pcre)

1.5

1.5
Option 3 - cap Pn , Pnd

Option 4 - cap Pn , transition Pnd

Pn/Py

Pn/Py

0.5

0.5

0.5

1.5

2.5

0.5

0.5

1.5
0.5

c =(Py /Pcre)

c =(Py /Pcre)

1.5

1.5
Option 5 - transition Pn (Option A), transition Pnd
l

Option 6 - transition Pn (Option B), transition Pnd


l
Pn/Py

Pn/Py

0.5

0.5

0.5

1.5

2.5

0.5

0.5

1.5

2
0.5

c =(Py /Pcre)

c =(Py /Pcre)

Figure8.21Comparisonofsimulatedteststrengthstopredictionsforcolumnswithglobalbucklingcontrolledfailures(i.e.,nolocalinteraction)asafunctionof
globalslenderness

348

1.5

Ptest/Pn

Ptest/Pn

1.5

0.5

Local Controlled
0.5

Option 1 - Py everywhere
0

0.5

Option 2 - Pynet everywhere


1.5

2.5

0.5

0.5

Ptest/Pn

Ptest/Pn

1.5

0.5

0.5

1.5

2.5

0.5

Ptest/Pn

Ptest/Pn

1.5

0.5

2.5

2.5

0.5

Option 5 - transition Pn (Option A), transition Pnd


l
1

l=(Pne/Pcrl)

1.5

0.5

1.5

1.5
0.5

l=(Pne/Pcrl)

Option 4 - cap Pn , transition Pnd


l

0.5

2.5

0.5

Option 3 - cap Pn , Pnd


l
0

l=(Pne/Pcrl)

1.5

1.5
0.5

l=(Pne/Pcrl)

Option 6 - transition Pn (Option B), transition Pnd


l
2.5

0.5

0.5

1.5

2
0.5

l=(Pne/Pcrl)

l=(Pne/Pcrl)

Figure8.22Testtopredictedratiosforlocalbucklingcontrolledsimulatedcolumnfailuresasafunctionoflocalslenderness

349

1.5

Ptest/Pn

Ptest/Pn

1.5

0.5

Distortional Controlled
0.5

Option 1 - Py everywhere
0

0.5

Option 2 - Pynet everywhere


1.5

2.5

0.5

0.5

Ptest/Pn

Ptest/Pn

1.5

0.5

0.5

1.5

2.5

0.5

Ptest/Pn

Ptest/Pn

1.5

0.5

2.5

2.5

0.5

Option 5 - transition Pn (Option A), transition Pnd


l
1

d=(Py /Pcrd)

1.5

0.5

1.5

1.5
0.5

d=(Py /Pcrd)

Option 4 - cap Pn , transition Pnd


l

0.5

2.5

0.5

Option 3 - cap Pn , Pnd


l
0

dnet=(Pynet/Pcrd)

1.5

1.5
0.5

d=(Py /Pcrd)

Option 6 - transition Pn (Option B), transition Pnd


l
2.5

0.5

0.5

1.5

2
0.5

d=(Py /Pcrd)

d=(Py /Pcrd)

Figure8.23Testtopredictedratiosfordistortionalbucklingcontrolledsimulatedcolumnfailuresasafunctionofdistortionalslenderness

350

1.5

Ptest/Pne

Ptest/Pne

1.5

0.5

0.5

Option 1 - Py everywhere
0

0.5

1.5

2.5

0.5

0.5

Ptest/Pne

Ptest/Pne

1.5

0.5

1.5

2.5

0.5

Ptest/Pne

Ptest/Pne

1.5

0.5

2.5

2.5

0.5

Option 5 - transition Pn (Option A), transition Pnd


l
1

c =(Py /Pcre)

1.5

0.5

1.5

1.5
0.5

c =(Py /Pcre)

Option 4 - cap Pn , transition Pnd


l

0.5

2.5

0.5

Option 3 - cap Pn , Pnd


l
0.5

cnet=(Pynet/Pcre)

1.5

1.5
0.5

c =(Py /Pcre)

Global Controlled

Option 2 - Pynet everywhere

Option 6 - transition Pn (Option B), transition Pnd


l
2.5

0.5

0.5

1.5

2
0.5

c =(Py /Pcre)

c =(Py /Pcre)

Figure8.24Testtopredictedratiosforsimulatedglobalbucklingcontrolledcolumnfailures(i.e.,nolocalbucklinginteraction)asafunctionofglobalslenderness

351

1.5

Ptest/Pn

Ptest/Pn

1.5

0.5

0.5

Option 1 - Py everywhere
0
0.5

0.6

0.7

0.8

0.9

0
0.5

0.6

0.7

1.5

1.5

0.5

0.7

0.8

0.9

0
0.5

0.6

0.7

1.5

Ptest/Pn

Ptest/Pn

0.8

0.9

0.9

Anet/Ag

1.5

0.5

0.5

Option 5 - transition Pn (Option A), transition Pnd

Option 6 - transition Pn (Option B), transition Pnd

0.6

Option 4 - cap Pn , transition Pnd


l

Anet/Ag

0
0.5

0.9

0.5

Option 3 - cap Pn , Pnd


l
0.6

0.8
Anet/Ag

Ptest/Pn

Ptest/Pn

Anet/Ag

0
0.5

Local Controlled

Option 2 - Pynet everywhere

0.7

0.8

0.9

0
0.5

Anet/Ag

0.6

0.7

0.8
Anet/Ag

Figure8.25Testtopredictedratiosforsimulatedlocalbucklingcontrolledcolumnfailuresasafunctionofnetcrosssectionalareatogrosscrosssectionalarea

352

1.5

Ptest/Pn

Ptest/Pn

1.5

0.5

0.5

Option 1 - Py everywhere
0
0.5

0.6

0.7

0.8

0.9

0
0.5

0.6

0.7

1.5

1.5

0.5

0.7

0.8

0.9

0
0.5

0.6

0.7

1.5

Ptest/Pn

Ptest/Pn

0.8

0.9

0.9

Anet/Ag

1.5

0.5

0.5

Option 5 - transition Pn (Option A), transition Pnd

Option 6 - transition Pn (Option B), transition Pnd

0.6

Option 4 - cap Pn , transition Pnd


l

Anet/Ag

0
0.5

0.9

0.5

Option 3 - cap Pn , Pnd


l
0.6

0.8
Anet/Ag

Ptest/Pn

Ptest/Pn

Anet/Ag

0
0.5

Distortional Controlled

Option 2 - Pynet everywhere

0.7

0.8

0.9

0
0.5

Anet/Ag

0.6

0.7

0.8
Anet/Ag

Figure8.26Testtopredictedratiosforsimulateddistortionalbucklingcontrolledcolumnfailuresasafunctionofnetcrosssectionalareatogrosscrosssectional
area

353

1.5

Ptest/Pn

Ptest/Pn

1.5

0.5

0.5

Option 1 - Py everywhere
0
0.5

0.6

0.7

0.8

0.9

0
0.5

0.6

0.7

1.5

1.5

0.5

0.7

0.8

0.9

0
0.5

0.6

0.7

1.5

Ptest/Pn

Ptest/Pn

0.8

0.9

0.9

Anet/Ag

1.5

0.5

0.5

Option 5 - transition Pn (Option A), transition Pnd

Option 6 - transition Pn (Option B), transition Pnd

0.6

Option 4 - cap Pn , transition Pnd


l

Anet/Ag

0
0.5

0.9

0.5

Option 3 - cap Pn , Pnd


l
0.6

0.8
Anet/Ag

Ptest/Pn

Ptest/Pn

Anet/Ag

0
0.5

Global Controlled

Option 2 - Pynet everywhere

0.7

0.8

0.9

0
0.5

Anet/Ag

0.6

0.7

0.8
Anet/Ag

Figure8.27Testtopredictedratiosforsimulatedglobalbucklingcontrolledcolumnfailuresasafunctionofnetcrosssectionalareatogrosscrosssectionalarea

354

1.5

Ptest/Pn

Ptest/Pn

1.5

0.5

Local Controlled

0.5

Option 1 - Py everywhere
0

10

15

20
L/h

25

30

35

40

1.5

1.5

Ptest/Pn

Ptest/Pn

Option 2 - Pynet everywhere

0.5

10

15

20
L/h

25

30

35

40

1.5

1.5

0.5

10

20
L/h

25

30

35

40

15

20
L/h

25

10

15

20
L/h

25

30

35

40

30

35

40

0.5

Option 5 - transition Pn (Option A), transition Pnd


l
0

15

Option 4 - cap Pn , transition Pnd


l

Ptest/Pn

Ptest/Pn

10

0.5

Option 3 - cap Pn , Pnd


l
0

Option 6 - transition Pn (Option B), transition Pnd


l
30

35

40

10

15

20
L/h

25

Figure8.28Testtopredictedratiosforsimulatedlocalbucklingcontrolledcolumnfailuresasafunctionofcolumnlength,L,toflatwebwidth,h

355

1.5

Ptest/Pn

Ptest/Pn

1.5

0.5

Distortional Controlled

0.5

Option 1 - Py everywhere
0

10

15

20
L/h

25

30

35

40

1.5

1.5

Ptest/Pn

Ptest/Pn

Option 2 - Pynet everywhere

0.5

10

15

20
L/h

25

30

35

40

1.5

1.5

0.5

10

20
L/h

25

30

35

40

15

20
L/h

25

10

15

20
L/h

25

30

35

40

30

35

40

0.5

Option 5 - transition Pn (Option A), transition Pnd


l
0

15

Option 4 - cap Pn , transition Pnd


l

Ptest/Pn

Ptest/Pn

10

0.5

Option 3 - cap Pn , Pnd


l
0

Option 6 - transition Pn (Option B), transition Pnd


l
30

35

40

10

15

20
L/h

25

Figure8.29Testtopredictedratiosforsimulateddistortionalbucklingcontrolledcolumnfailuresasafunctionofcolumnlength,L,toflatwebwidth,h

356

1.5

Ptest/Pn

Ptest/Pn

1.5

0.5

Global Controlled
0.5

Option 1 - Py everywhere
0

10

15

20
L/h

25

30

35

40

1.5

1.5

Ptest/Pn

Ptest/Pn

Option 2 - Pynet everywhere

0.5

10

15

20
L/h

25

30

35

40

1.5

1.5

0.5

10

20
L/h

25

30

35

40

15

20
L/h

25

10

15

20
L/h

25

30

35

40

30

35

40

0.5

Option 5 - transition Pn (Option A), transition Pnd


l
0

15

Option 4 - cap Pn , transition Pnd


l

Ptest/Pn

Ptest/Pn

10

0.5

Option 3 - cap Pn , Pnd


l
0

Option 6 - transition Pn (Option B), transition Pnd


l
30

35

40

10

15

20
L/h

25

Figure8.30Testtopredictedratiosforsimulatedglobalbucklingcontrolledcolumnfailuresasafunctionofcolumnlength,L,towebwidth,h

357

1.7 DSM comparison to experimental column database


109B

The six DSM options are now compared to the column experiment database first

assembled in Section 4.2.6.2. The database contains the elastic buckling properties of
1756H

eachcolumn,includingthepresenceofholesandtheinfluenceofboundaryconditions,
aswellasthetestedstrengths. Figure8.31through Figure8.34comparetheexperiment
175H

1758H

strengths to DSM predictions for local, distortional, and global buckling controlled
column failures. Option 1 is observed to be an accurate predictor of column strength
when local, distortional, and global slenderness are high, but overpredicts the tested
strength as slenderness decreases. This trend is consistent with the simulated
experimentcomparisoninSection 8.1.6andemphasizestheneedforalimitoncolumn
1759H

strength when yielding at the net section controls the failure of a column with holes.
Option 2 is even more conservative in this study when compared to the simulated
column study because the tested specimens considered only have one hole, and
therefore employing Pynet produces unrealistically high column slenderness. Option 3
shifts column specimens from the global buckling failure group to the local buckling
failuregroupwiththePynetlimitonPnl,resultinginimprovedaccuracywhencompared
toOption2.FourcolumnsintheOption3distortionalbucklingfailuregrouparestill
overpredicted by more than 10% though as observed in Figure 8.33. Option 4 and
1760H

Option 5 improve the accuracy of the underpredicted specimen strengths in Option 3


withtheadditionofthedistortionalandlocalyieldcontroltransitionstothenetsection.
Option6isanoverlyconservativepredictorofcolumnsfailingbyglobalbuckling.

358

Table 8.2 summarizes the testtopredicted ratio statistics for all columns in the
176H

database. Options 3, 4, and 5 are identified as the methods with the mean closest to
unityandwiththeloweststandarddeviations.Thestatisticsforjustthestubcolumns
(c<0.20) in Table 8.3 confirm the viability of DSM Options 3, 4, and 5, and provides
1762H

more direct evidence that holes limit the column strength to the net section Pynet; the
mean testtopredicted ratio is 0.84 for global (yielding) failures of stub columns
employingOption1.

Table8.2DSMtesttopredictedratiostatisticsforcolumnexperiments
Option
1

Description
Py everywhere

Mean
1.03

Local buckling
SD

0.11
0.87

# of tests
52

Mean
1.09

Distortional buckling
SD
# of tests

0.16
0.83
15

Mean
1.06

Global buckling
SD

0.17
0.82

# of tests
11

Pynet everywhere

1.17

0.09

0.89

47

1.22

0.13

0.87

15

1.17

0.15

0.85

16

Cap Pnl, Pnd

1.07

0.08

0.90

42

1.06

0.13

0.85

29

1.16

0.09

0.90

Transition Pnd, Cap Pnl

1.07

0.08

0.90

40

1.10

0.11

0.87

33

1.19

0.08

0.90

Transition Pnd and Pnl (Option A)

1.06

0.08

0.89

47

1.13

0.10

0.89

26

1.19

0.08

0.90

Transition Pnd and Pnl (Option B)

1.12

0.15

0.84

56

1.14

0.10

0.89

22

---

---

---

Table8.3DSMtesttopredictedratiostatisticsforcolumnexperiments(stubcolumnsonly)
Option
1

Description
Py everywhere

Mean
0.98

Local buckling
SD

0.10
0.88

# of tests
33

Mean
0.83

Distortional buckling
SD
# of tests

0.01
0.92
3

Mean
0.84

Global buckling
SD

0.08
0.88

# of tests
3

Pynet everywhere

1.12

0.07

0.90

28

1.03

0.06

0.91

1.07

0.12

0.86

Cap Pnl, Pnd

1.03

0.06

0.91

23

1.00

0.12

0.86

16

---

---

---

Transition Pnd, Cap Pnl

1.04

0.06

0.91

21

1.06

0.11

0.87

18

---

---

---

Transition Pnd and Pnl (Option A)

1.03

0.07

0.90

28

1.11

0.10

0.88

11

---

---

---

Transition Pnd and Pnl (Option B)

1.03

0.07

0.90

29

1.11

0.10

0.88

10

---

---

---

359

1.5

1.5
Option 1 - Py everywhere

Option 2 - Pynet everywhere


1
Pn/Pne

Pn/Pne

0.5

0.5

1.5

2.5

0.5

0.5

1.5

0.5

Option 4 - cap Pn , transition Pnd


l

DSM Local Prediction


Local Controlled Test

1
Pn/Pne

Pn/Pne

0.5

0.5

1.5

2.5

0.5

0.5

1.5

0.5

2.5

2.5

0.5

l=(Pne/Pcrl)

l=(Pne/Pcrl)

1.5

1.5
Option 5 - transition Pn (Option A), transition Pnd

Option 6 - transition Pn (Option B), transition Pnd

1
Pn/Pne

1
Pn/Pne

1.5
Option 3 - cap Pn , Pnd
l

0.5

2.5

l=(Pne/Pcrl)

1.5

2
0.5

l=(Pne/Pcrl)

0.5

1.5

2.5

0.5

0.5

0.5

1.5

2
0.5

l=(Pne/Pcrl)

l=(Pne/Pcrl)

Figure8.31Comparisonofexperimentalteststrengthstopredictionsforcolumnswithlocalbucklingcontrolledfailuresasafunctionoflocalslenderness(tested
strengthisnormalizedbyPne)

360

1.5

1.5
Option 1 - Py everywhere

Option 2 - Pynet everywhere

Pn/Py

Pn/Py

0.5

0.5

0.5

1.5

2.5

0.5

0.5

1.5

2.5

0.5

l=(Pne/Pcrl)

l=(Pne/Pcrl)

1.5

1.5
Option 3 - cap Pn , Pnd

Option 4 - cap Pn , transition Pnd

DSM Local Prediction


Local Controlled Test

Pn/Py

Pn/Py

0.5

0.5

0.5

1.5

2.5

0.5

0.5

1.5

2.5

2.5

0.5

l=(Pne/Pcrl)

l=(Pne/Pcrl)

1.5

1.5
Option 5 - transition Pn (Option A), transition Pnd
l

Option 6 - transition Pn (Option B), transition Pnd


l
Pn/Py

Pn/Py

0.5

0.5

0.5

1.5

2.5

0.5

0.5

1.5

2
0.5

l=(Pne/Pcrl)

l=(Pne/Pcrl)

Figure8.32Comparisonofexperimentalteststrengthstopredictionsforcolumnswithlocalbucklingcontrolledfailuresasafunctionoflocalslenderness(tested
strengthisnormalizedbyPy)

361

1.5

1.5
Option 1 - Py everywhere

Option 2 - Pynet everywhere

Pn/Py

Pn/Py

DSM Dist. Prediction


Dist. Controlled Test

0.5

0.5

0.5

1.5

2.5

0.5

0.5

1.5

2.5

2.5

2.5

0.5

d=(Py /Pcrd)

dnet=(Pynet/Pcrd)

1.5

1.5
Option 3 - cap Pn , Pnd

Option 4 - cap Pn , transition Pnd

Pn/Py

Pn/Py

0.5

0.5

0.5

1.5

2.5

0.5

0.5

1.5
0.5

d=(Py /Pcrd)

d=(Py /Pcrd)

1.5

1.5
Option 5 - transition Pn (Option A), transition Pnd
l

Option 6 - transition Pn (Option B), transition Pnd


l
Pn/Py

Pn/Py

0.5

0.5

0.5

1.5

2.5

0.5

0.5

1.5

2
0.5

d=(Py /Pcrd)

d=(Py /Pcrd)

Figure8.33Comparisonofexperimentalteststrengthstopredictionsforcolumnswithdistortionalbucklingcontrolledfailuresasafunctionofdistortional
slenderness

362

1.5

1.5
Option 2 - Pynet everywhere

Option 1 - Py everywhere

Pn/Py

Pn/Py

DSM Global Prediction


Global Controlled Test

0.5

0.5

0.5

1.5

2.5

0.5

0.5

1.5

2.5

2.5

2.5

0.5

c =(Py /Pcre)

cnet=(Pynet/Pcre)

1.5

1.5
Option 3 - cap Pn , Pnd

Option 4 - cap Pn , transition Pnd

Pn/Py

Pn/Py

0.5

0.5

0.5

1.5

2.5

0.5

0.5

1.5
0.5

c =(Py /Pcre)

c =(Py /Pcre)

1.5

1.5
Option 5 - transition Pn (Option A), transition Pnd
l

Option 6 - transition Pn (Option B), transition Pnd


l
Pn/Py

Pn/Py

0.5

0.5

0.5

1.5

2.5

0.5

0.5

1.5

2
0.5

c =(Py /Pcre)

c =(Py /Pcre)

Figure8.34Comparisonofexperimentalteststrengthstopredictionsforcolumnswithglobalbucklingcontrolledfailuresasafunctionofglobalslenderness

363

1.5

Ptest/Pn

Ptest/Pn

1.5

0.5

Local Controlled
0.5

Option 1 - Py everywhere
0

0.5

Option 2 - Pynet everywhere


1.5

2.5

0.5

0.5

Ptest/Pn

Ptest/Pn

1.5

0.5

0.5

1.5

2.5

0.5

Ptest/Pn

Ptest/Pn

1.5

0.5

2.5

2.5

0.5

Option 5 - transition Pn (Option A), transition Pnd


l
1

l=(Pne/Pcrl)

1.5

0.5

1.5

1.5
0.5

l=(Pne/Pcrl)

Option 4 - cap Pn , transition Pnd


l

0.5

2.5

0.5

Option 3 - cap Pn , Pnd


l
0

l=(Pne/Pcrl)

1.5

1.5
0.5

l=(Pne/Pcrl)

Option 6 - transition Pn (Option B), transition Pnd


l
2.5

0.5

0.5

1.5

2
0.5

l=(Pne/Pcrl)

l=(Pne/Pcrl)

Figure8.35Testtopredictedratiosforexperimentlocalbucklingcontrolledcolumnfailuresasafunctionoflocalslenderness

364

1.5

Ptest/Pn

Ptest/Pn

1.5

0.5

Distortional Controlled
0.5

Option 1 - Py everywhere
0

0.5

Option 2 - Pynet everywhere


1.5

2.5

0.5

0.5

Ptest/Pn

Ptest/Pn

1.5

0.5

0.5

1.5

2.5

0.5

Ptest/Pn

Ptest/Pn

1.5

0.5

2.5

2.5

0.5

Option 5 - transition Pn (Option A), transition Pnd


l
1

d=(Py /Pcrd)

1.5

0.5

1.5

1.5
0.5

d=(Py /Pcrd)

Option 4 - cap Pn , transition Pnd


l

0.5

2.5

0.5

Option 3 - cap Pn , Pnd


l
0

dnet=(Pynet/Pcrd)

1.5

1.5
0.5

d=(Py /Pcrd)

Option 6 - transition Pn (Option B), transition Pnd


l
2.5

0.5

0.5

1.5

2
0.5

d=(Py /Pcrd)

d=(Py /Pcrd)

Figure8.36Testtopredictedratiosforexperimentdistortionalbucklingcontrolledcolumnfailuresasafunctionofdistortionalslenderness

365

1.5

Ptest/Pne

Ptest/Pne

1.5

0.5

Global Controlled

0.5

Option 1 - Py everywhere
0

0.5

Option 2 - Pynet everywhere


1.5

2.5

0.5

0.5

Ptest/Pne

Ptest/Pne

1.5

0.5

0.5

1.5

2.5

0.5

Ptest/Pne

Ptest/Pne

1.5

0.5

2.5

2.5

0.5

Option 5 - transition Pn (Option A), transition Pnd


l
1

c =(Py /Pcre)

1.5

0.5

1.5

1.5
0.5

c =(Py /Pcre)

Option 4 - cap Pn , transition Pnd


l

0.5

2.5

0.5

Option 3 - cap Pn , Pnd


l
0

cnet=(Pynet/Pcre)

1.5

1.5
0.5

c =(Py /Pcre)

Option 6 - transition Pn (Option B), transition Pnd


l
2.5

0.5

0.5

1.5

2
0.5

c =(Py /Pcre)

c =(Py /Pcre)

Figure8.37Testtopredictedratiosforexperimentglobalbucklingcontrolledcolumnfailuresasafunctionofglobalslenderness

366

1.5

Ptest/Pn

Ptest/Pn

1.5

0.5

Local Controlled
0.5

Option 1 - Py everywhere
0
0.5

0.6

Option 2 - Pynet everywhere

0.7

0.8

0.9

0
0.5

0.6

0.7

1.5

1.5

0.5

0.6

0.7

0.8

0.9

0
0.5

0.6

0.7

1.5

Ptest/Pn

Ptest/Pn

0.8

0.9

0.9

Anet/Ag

1.5

0.5

0.5

Option 5 - transition Pn (Option A), transition Pnd

Option 6 - transition Pn (Option B), transition Pnd

0.6

Option 4 - cap Pn , transition Pnd


l

Anet/Ag

0
0.5

0.9

0.5

Option 3 - cap Pn , Pnd


l
0
0.5

0.8
Anet/Ag

Ptest/Pn

Ptest/Pn

Anet/Ag

0.7

0.8

0.9

0
0.5

Anet/Ag

0.6

0.7

0.8
Anet/Ag

Figure8.38Testtopredictedratiosforexperimentlocalbucklingcontrolledcolumnfailuresasafunctionofnetcrosssectionalareatogrosscrosssectionalarea

367

1.5

Ptest/Pn

Ptest/Pn

1.5

0.5

Distortional Controlled
0.5

Option 1 - Py everywhere
0
0.5

0.6

Option 2 - Pynet everywhere

0.7

0.8

0.9

0
0.5

0.6

0.7

1.5

1.5

0.5

0.6

0.7

0.8

0.9

0
0.5

0.6

0.7

1.5

Ptest/Pn

Ptest/Pn

0.8

0.9

0.9

Anet/Ag

1.5

0.5

0.5

Option 5 - transition Pn (Option A), transition Pnd

Option 6 - transition Pn (Option B), transition Pnd

0.6

Option 4 - cap Pn , transition Pnd


l

Anet/Ag

0
0.5

0.9

0.5

Option 3 - cap Pn , Pnd


l
0
0.5

0.8
Anet/Ag

Ptest/Pn

Ptest/Pn

Anet/Ag

0.7

0.8

0.9

0
0.5

Anet/Ag

0.6

0.7

0.8
Anet/Ag

Figure8.39Testtopredictedratiosforexperimentdistortionalbucklingcontrolledcolumnfailuresasafunctionofnetcrosssectionalareatogrosscrosssectional
area

368

1.5

Ptest/Pn

Ptest/Pn

1.5

0.5

0.5

Option 1 - Py everywhere
0
0.5

0.6

0.7

0.8

0.9

0
0.5

0.6

0.7

1.5

1.5

0.5

0.7

0.8

0.9

0
0.5

0.6

0.7

1.5

Ptest/Pn

Ptest/Pn

0.8

0.9

0.9

Anet/Ag

1.5

0.5

0.5

Option 5 - transition Pn (Option A), transition Pnd

Option 6 - transition Pn (Option B), transition Pnd

0.6

Option 4 - cap Pn , transition Pnd


l

Anet/Ag

0
0.5

0.9

0.5

Option 3 - cap Pn , Pnd


l
0.6

0.8
Anet/Ag

Ptest/Pn

Ptest/Pn

Anet/Ag

0
0.5

Global Controlled

Option 2 - Pynet everywhere

0.7

0.8

0.9

0
0.5

Anet/Ag

0.6

0.7

0.8
Anet/Ag

Figure8.40Testtopredictedratiosforexperimentglobalbucklingcontrolledcolumnfailuresasafunctionofnetcrosssectionalareatogrosscrosssectionalarea

369

1.5

Ptest/Pn

Ptest/Pn

1.5

0.5

Local Controlled

0.5

Option 1 - Py everywhere
0

10

15

20
L/h

25

30

35

40

1.5

1.5

Ptest/Pn

Ptest/Pn

Option 2 - Pynet everywhere

0.5

10

15

20
L/h

25

30

35

40

1.5

1.5

0.5

10

20
L/h

25

30

35

40

15

20
L/h

25

10

15

20
L/h

25

30

35

40

30

35

40

0.5

Option 5 - transition Pn (Option A), transition Pnd


l
0

15

Option 4 - cap Pn , transition Pnd


l

Ptest/Pn

Ptest/Pn

10

0.5

Option 3 - cap Pn , Pnd


l
0

Option 6 - transition Pn (Option B), transition Pnd


l
30

35

40

10

15

20
L/h

25

Figure8.41Testtopredictedratiosforexperimentlocalbucklingcontrolledcolumnfailuresasafunctionofcolumnlength,L,toflatwebwidth,h

370

1.5

Ptest/Pn

Ptest/Pn

1.5

0.5

Distortional Controlled

0.5

Option 1 - Py everywhere
0

10

15

20
L/h

25

30

35

40

1.5

1.5

Ptest/Pn

Ptest/Pn

Option 2 - Pynet everywhere

0.5

10

15

20
L/h

25

30

35

40

1.5

1.5

0.5

10

20
L/h

25

30

35

40

15

20
L/h

25

10

15

20
L/h

25

30

35

40

30

35

40

0.5

Option 5 - transition Pn (Option A), transition Pnd


l
0

15

Option 4 - cap Pn , transition Pnd


l

Ptest/Pn

Ptest/Pn

10

0.5

Option 3 - cap Pn , Pnd


l
0

Option 6 - transition Pn (Option B), transition Pnd


l
30

35

40

10

15

20
L/h

25

Figure8.42Testtopredictedratiosforexperimentdistortionalbucklingcontrolledcolumnfailuresasafunctionofcolumnlength,L,toflatwebwidth,h

371

1.5

Ptest/Pn

Ptest/Pn

1.5

0.5

Global Controlled

0.5

Option 1 - Py everywhere
0

10

15

20
L/h

25

30

35

40

1.5

1.5

Ptest/Pn

Ptest/Pn

Option 2 - Pynet everywhere

0.5

10

15

20
L/h

25

30

35

40

1.5

1.5

0.5

10

20
L/h

25

30

35

40

15

20
L/h

25

10

15

20
L/h

25

30

35

40

30

35

40

0.5

Option 5 - transition Pn (Option A), transition Pnd


l
0

15

Option 4 - cap Pn , transition Pnd


l

Ptest/Pn

Ptest/Pn

10

0.5

Option 3 - cap Pn , Pnd


l
0

Option 6 - transition Pn (Option B), transition Pnd


l
30

35

40

10

15

20
L/h

25

Figure8.43Testtopredictedratiosforexperimentglobalbucklingcontrolledcolumnfailuresasafunctionofcolumnlength,L,towebwidth,h

372

1.8 Recommendations DSM for columns with holes


10B

Options3,4,and5arepresentedasviableproposalsforextendingDSMtocolumns

with holes. This recommendation is based on the testtopredicted statistics and data
trends presented in Section 8.1.6 and Section 8.1.7, and also considers the effort to
1763H

1764H

implement the modifications and their ease of use by design engineers. Option 3
accountsforthereductionincolumnstrengthfromthepresenceofholesbycappingPnl
and Pnd at Pynet. This is a simple modification to implement in the Specification and
avoids additional calculation work for a design engineer (except for that required to
calculatethecriticalelasticbucklingloadsincludingtheinfluenceoftheholes).Options
4and5arerefinementsofOption3,wherethecaponPnlandPndbecomesatransition
from an elastic buckling controlled failure mode to a yield controlled failure at Pynet.
These two methods require additional effort from the designer when compared to
Option3,buttheyhaveanimportantadvantage.Options4and5aremorecloselytied
to the failure mechanisms influencing column strength because they capture the yield
transitiontothenetsectionintheirpredictions.Thetransitionsincreasetheprobability
that strength will be accurately predicted for general column and hole geometries.
Option5hastheadditionaladvantageofcapturingtheinfluenceofayieldtransitionfor
closed crosssections that do not experience distortional buckling. This generality is
whatmotivatestheuseoftheDirectStrengthMethod(AISIS1002007,Appendix1).

373

8.2 DSMforlaterallybracedbeamswithholes
49B

2.1 Database of simulated column experiments


1B

Simulatedexperimentswereconductedon125Csectionlaterallybracedbeamswith

evenlyspacedcircularwebholesinABAQUS.Crosssectionswerespecificallyselected
withcustomMatlabcodeemployingtheexistingDSMdesigncurvestoidentifybeams
predisposed to local and distortional bucklingcontrolled failures. The crosssections
were chosen from a catalog of 99 industry standard Csections published by the Steel
Stud Manufacturers Association (SSMA 2001). The nominal outtoout dimensions
provided in the SSMA catalog were converted to centerline dimensions and then
constructedinABAQUSwiththemeshingproceduredescribedinSection 7.2.1.1.The
1765H

beams in the database have a constant length L=48 in. to accommodate multiple local
anddistortionalbucklinghalfwavesalongthebeam.Evenlyspacedcircularwebholes
wereplacedinthecolumnswithholespacingS(definedin Figure3.2)of16inches(i.e.,
176H

threeevenlyspacedholes).Theholeswerecenteredtransverselyinthewebandtheir
depth(diameter),hhole,wasvariedsuchthattheratioofthenetmomentofinertia,Inet,to
thegrosscrosssectionalarea,Ig,rangedbetween0.85and1.0.

TheABAQUS boundaryconditionsandapplicationofloading,describedin Figure


176H

8.44, are implemented to be consistent with CUFSM, i.e. pinnedpinned and freeto
warpwithauniformstressappliedatthememberends.Eachbeamislaterallybraced
by restraining the compression flange at the midlength of the beam. (Initial modeling
trials, where all nodes centered in the compression flange were laterally restrained,

374

resulted in simulated strengths 25% higher than DSM predictions for beams without
holes.)Consistentnodalloadswereappliedtosimulatethelinearstressgradientatthe
beam ends (see Section 7.2.1.2 for information on S9R5 consistent nodal loads). The
1768H

loads (a reference moment of 1 kipin. was applied at each end in ABAQUS) were
distributedoverthefirsttwosetsofcrosssectionnodestoavoidlocalizedfailuresatthe
loadededges.
End cross-section nodes
restrained in 2 and 3
Node centered in compression flange at
longitudinal midline restrained in 1 (to prevent
rigid body motion) and 3 (for laterally bracing)

End cross-section nodes


restrained in 2 and 3

2
5

4
1

Moment applied as consistent nodal loads


over two sets of cross-section nodes to
avoid edge failures (Typ.)

Figure8.44ABAQUSsimulatedbeamexperimentsboundaryconditionsandapplicationofloading

The ABAQUS simulations were performed with the modified Riks nonlinear

solution algorithm. Automatic time stepping was enabled with a suggested initial arc
length step of 1 (the Riks method increments in units of energy, in this case kipin.), a
maximum step size of 3, and the maximum number of solution increments set at 300.
Metal plasticity was simulated with the material modeling procedure described in
Section7.2.1.4.Theplastictruestressstraincurveforspecimen362148HinAppendix
1769H

170H

H was assumed for all column models (but modified such that plasticity starts at the
yieldstress,seeSection7.2.1.4),wherethesteelyieldstressFy=58.6ksi.Residualstresses
17H

375

and initial plastic strains, as discussed in Section 7.2.1.6, were not considered in the
172H

ABAQUS models because their implementation requires further validation and they
werenotobservedtomarkedlyinfluencecolumnultimatestrength(see Figure7.48and
173H

Figure7.49).
174H

ImperfectionswereimposedontheinitialbeamgeometryinABAQUSwithcustom

Matlab code which combines the local and distortional buckling crosssection mode
shapes from CUFSM along the column length. Two simulations were performed for
each beam, one model with 25% CDF local and distortional imperfection magnitudes
andtheothermodelwith75%CDFlocalanddistortionalimperfectionmagnitudes(see
Section7.2.1.5forlocalanddistortionalimperfectiondefinitions).
175H

The local (Mcrl) and distortional (Mcrd) critical elastic buckling loads were predicted

for each beam with custom Matlab code based on the CUFSM prediction methods
describedinSection4.3.Thedatabaseofsimulatedbeamexperiments,includingcross
176H

section type, column and hole geometry, simulated ultimate strength (Mtest25 and Mtest75)
and critical elastic buckling loads for each beam (including the presence of holes) is
providedinAppendixL.
17H

2.2 Local buckling study


12B

Twelvebeamsfromthesimulationdatabasein Appendix Lwerechosentostudy


178H

the influence of web holes on the ultimate strength of laterally braced beams
predisposedtoalocalbucklingcontrolledfailure.ThebeamshaveSSMAcrosssections
whichresultinalocalbucklingslenderness,l,rangingfrom1.3to2.0.(Theslenderness

376

range considered here is relatively narrow because only 12 of the 99 SSMA cross
sections, when employed as laterally braced beams, are controlled by a local buckling
failure. The majority of beam crosssections are predicted to exhibit a distortional
bucklingcontrolled failure.) The web of each beam contains three evenly spaced
circularholeswheretheholespacingS=16in.Theholedepth(diameter),hhole,isvaried
foreachbeamtoproduceInet/Igof1.0(noholes),0.95,0.90,and0.85.RefertoAppendix
179H

L,StudyTypeL,forspecificbeamcrosssectionandholegeometryinformation.

The simulation results for Inet/Ig =1.0, 0.95, 0.90, and 0.85, are compared to the DSM

distortionalbucklingpredictioncurveinFigure8.46toFigure8.49.Thebeamstrengths,
1780H

178H

Mtest25 and Mtest75, without holes (Inet/Ig =1.0) are consistent with the DSM design curve as
shownin Figure8.46a,confirmingthatthenonlinearsimulationprotocoldevelopedfor
1782H

columns in Section 7.2 is also viable when conducting coldformed steel beam
1783H

simulations.Themeanandstandarddeviationofthesimulatedtesttopredictedratiois
1.05 and 0.05 respectively for 25% CDF local and distortional imperfections, and 1.03
and 0.05 for 75% CDF local and distortional imperfections. For the beams with holes,
thesimulatedteststrengthsdivergefromtheDSMpredictioncurveaslocalslenderness,

l=(Myg/Mcrl)0.5, decreases as shown in Figure 8.47a to Figure 8.49a (Myg is the yield
1784H

1785H

momentofthecolumncalculatedwiththegrosscrosssectionalareaIg).Thisdivergent
trend in Mtest with decreasing l is consistent with the column results with holes
discussedinSection 8.1,whereelasticbucklingcontrolledthefailurewhenslenderness
1786H

was high and transitioned to yielding and collapse of the net section as slenderness
decreased. Figure8.45showstheloaddeformationresponseatultimatelimitstatefor
178H

377

anSSMA800S16243beamconsideredinthisstudy,andhighlightsthetransitionfrom
an elastic buckling controlledfailure to a yield controlledfailure at the net section as
holesizeincreases.
Elastic buckling controlled failure

Yielding and collapse of net section

Inet/Ig

1.0

0.95

0.90

Mtest25/Myg
l

0.72

0.68

0.64

0.85
0.61

1.45

1.63

1.45

1.45

Figure8.45SSMA800S16243beamwithwebholesconsideredintheDSMlocalbucklingstudy

Two modification options are proposed for the DSM local buckling beam design

curve:
Local Buckling (Option A)
The nominal flexural strength, Mnl, for local buckling shall be calculated in accordance with the following:
(a) For l l1

Mnl = Mne Mynet

(caponbeamstrength)

(b) For l1<ll2

l 2 l l1
Mnl = M

ynet M ynet M l 2

l l 2 l1

(c) For l > l2

Mnl = 1 0.15 M crl

where
l

l1

M ne

0.4

M
crl
M ne

(nonlinearyieldtransitionwhenMynet/Mne1)

0.4

M ne

(DSMlocalbucklingcurve,unchanged)

= M ne M crl
= 0.776( M ynet M ne ) 0.776

l2 = 0.776 2.4(M ynet M ne )3.5 1.4 , Mynet/Mne1


= 0.776, Mynet/Mne>1

l2 = 1 0.15(1 l 2 )0.8 (1 l 2 )0.8 M ne


Mynet = SfnetFy0.80My

(notransitionwhenMynet/Mne>1)
(limitreductionofthenetsectionto0.8My)

Sfnet = Section modulus at the hole(s) referenced to the extreme fiber at first yield
Mcrl = Critical elastic local beam buckling load including hole(s)

378

Local Buckling (Option B)


The nominal axial strength, Mnl, for local buckling shall be calculated in accordance with the following:
(a) For l l1

Mnl = Mynet (Mne /My)

(caponcolumnstrength)

(b) For l1<ll2

Mnl = M M ne M M ne M l 2 l l1
ynet
ynet
l2

M
y

M
y

(c) For l > l2

Mnl = 1 0.15 M crl

where
l

l1

M ne

0.4

M
crl
M ne

l l 2 l1

0.4

M ne

(DSMlocalbucklingcurve,unchanged)

= M ne M crl
= 0.776( M ynet M y ) 0.776

l2 = 0.776 2.4(M ynet M y )3.5 1.4

(nonlinearyieldtransition)

l2 = 1 0.15(1 l 2 )0.8 (1 l 2 )0.8 M ne


Mynet = SfnetFy0.80My

(limitreductionofthenetsectionto0.8My)

Sfnet = Section modulus at the hole(s) referenced to the extreme fiber at first yield
Mcrl = Critical elastic local beam buckling load including hole(s)

TheframeworkforOptionAandOptionBisbasedontheproposedmodifications

to the DSM local buckling column design curve presented in Section 8.1.4. Option A
178H

imposes a transition from the DSM local buckling curve to the net section limit, Mynet,
whenMynet<Mne.WhenMynet>Mne,OptionAassumesthatholesinfluenceonlythecritical
elastic buckling loads (Mcrl, Mcre) but otherwise do not change the failure mode of the
beam.OptionBalsoimposesatransitiontothenetbeamstrengthfromtheDSMlocal
failurecurve,althoughinthiscasetheyieldtransitionoccursforallvaluesofMynet/Mne.
Theproposedtransitionfromtheelasticbucklingfailureregimetotheyieldplateauis
nonlinearforbothOptionsAandBasdemonstratedin Figure8.47ato Figure8.49a,in
1789H

1790H

contrast to the linear transition for coldformed steel columns with holes (see Section
8.1.4).
179H

379

Allbeamsconsideredinthisstudyarelaterallybraced,i.e.global(lateraltorsional)

bucklingdoesnotinfluencebeamstrength,andthereforeOptionAandBwillproduce
thesamestrengthpredictions.Thevalidityofbothoptionsforlaterallybracedbeamsis
evaluatedinthefollowingsectionwiththesimulationdatabasein AppendixLandthe
1792H

experiment database assembled in Chapter 4. Future work is planned to evaluate


1793H

Option A and B for unbraced coldformed steel beams with holes, where lateral
torsionalbucklinginfluencesbeamstrength.
1.4

1.4
DSM (no holes)
FE 25% CDF imperfections
FE 75% CDF imperfections

0.8

0.8

0.6

0.6

0.4

0.4

0.2

0.2

0.5

1.5

2.5

local slenderness, =(My /Mcr )0.5


l
l

DSM (no holes)


DSM (proposed, with holes)
FE 25% CDF imperfections
FE 75% CDF imperfections

1.2

Mtest/Mn

Mtest/Mn

1.2

0.5

1.5

2.5

local slenderness, =(My /Mcr )0.5


l
l

Figure8.46Comparisonofsimulatedbeamstrengths(Inet/Ig=1.0,noholes)to(a)theexistingDSMlocal
bucklingdesigncurveandto(b)theproposedDSMlocalbucklingcurveforbeamswithholes
1.4

1.4
DSM (no holes)
FE 25% CDF imperfections
FE 75% CDF imperfections

0.8

0.8

0.6

0.6

0.4

0.4

0.2

0.2

0.5

1.5

local slenderness, =(My /Mcr )0.5


l
l

2.5

DSM (no holes)


DSM (proposed, with holes)
FE 25% CDF imperfections
FE 75% CDF imperfections

1.2

Mtest/Mn

Mtest/Mn

1.2

0.5

1.5

2.5

local slenderness, =(My /Mcr )0.5


l
l

Figure8.47Comparisonofsimulatedbeamstrengths(Inet/Ig=0.95)to(a)theexistingDSMlocalbuckling
designcurveandto(b)theproposedDSMlocalbucklingcurveforbeamswithholes

380

1.4

1.4
DSM (no holes)
FE 25% CDF imperfections
FE 75% CDF imperfections

0.8

0.8

0.6

0.6

0.4

0.4

0.2

0.2

0.5

1.5

2.5

DSM (no holes)


DSM (proposed, with holes)
FE 25% CDF imperfections
FE 75% CDF imperfections

1.2

Mtest/Mn

Mtest/Mn

1.2

0.5

local slenderness, =(My /Mcr )0.5


l
l

1.5

2.5

local slenderness, =(My /Mcr )0.5


l
l

Figure8.48Comparisonofsimulatedbeamstrengths(Inet/Ig=0.90)to(a)theexistingDSMlocalbuckling
designcurveandto(b)theproposedDSMlocalbucklingcurveforbeamswithholes
1.4

1.4
DSM (no holes)
FE 25% CDF imperfections
FE 75% CDF imperfections

0.8

0.8

0.6

0.6

0.4

0.4

0.2

0.2

0.5

1.5

2.5

DSM (no holes)


DSM (proposed, with holes)
FE 25% CDF imperfections
FE 75% CDF imperfections

1.2

Mtest/Mn

Mtest/Mn

1.2

local slenderness, =(My /Mcr )0.5


l
l

0.5

1.5

local slenderness, =(My /Mcr )0.5


l
l

2.5

Figure8.49Comparisonofsimulatedbeamstrengths(Inet/Ig=0.85)to(a)theexistingDSMlocalbuckling
designcurveandto(b)theproposedDSMlocalbucklingcurveforbeamswithholes

2.3 Distortional buckling study


13B

A group of 11 beams from the SSMA beam simulation database was chosen to

evaluate the influence of the ratio Inet/Ig on the tested strength of beams predicted to
collapse with a distortional failure mode. (Ig is the gross moment of inertia of a beam
andInetisthemomentofinertiaatthelocationofahole.)ThebeamshaveSSMAcross
sectionswhichresultinadistortionalbucklingslenderness, d,rangingfrom0.6to1.6.
(AllSSMAcrosssections,employedasbeamsandcontrolledbyadistortionalbuckling
failure,liewithinthisslendernessrange.)Inthisstudythebeamdepthsrangefrom4in.

381

to12in.ThewebofeachbeamhasthreecircularholeswheretheholespacingS=16in
(see Figure3.2forthedefinitionofS).Theholedepth(diameter),hhole,isvariedforeach
1794H

beamtoproduceInet/Igof1.0(noholes),0.95,and0.90.RefertoAppendixL,StudyType
1795H

D,forspecificcrosssectionandholegeometryinformation.

The simulation results for Inet/Ig =1.0, 0.95, and 0.90 are compared to the DSM

distortionalbucklingpredictioncurveinFigure8.51toFigure8.53.Thebeamstrengths,
1796H

179H

Mtest25 and Mtest75, without holes (Inet/Ig =1.0) are consistent with the DSM distortional
buckling design curve as shown in Figure 8.51a, with a trend of increasingly
1798H

conservativepredictionsasdistortionalslendernessincreases.Themeanandstandard
deviation of the simulated test to predicted ratio is 1.08 and 0.08 respectively for 25%
CDFlocalanddistortionalimperfections,and1.02and0.12for75%CDFimperfections.
Forthebeamswithholes,thesimulatedteststrengthsdemonstrateaslightdivergence
fromtheDSMpredictioncurveasdistortionalslenderness, d=(Myg/Mcrd)0.5,decreasesas
shownin Figure8.52aand Figure8.53a(Mygistheyieldmomentofthebeamcalculated
179H

180H

with the gross crosssectional area Ig). (Figure 8.52a and Figure 8.53a also demonstrate
180H

1802H

that Mcrd, predicted with the simplified method in Section 4.3.2.2, increases distortional
1803H

slendernessandshiftsthesimulateddataoffofthepredictioncurve.Futureresearchis
plannedtoimprovetheaccuracyofthissimplifiedmethod.)Thisdivergenttrendin
MtestwasalsoobservedinthelocalbucklingcontrolledbeamstudyinSection 8.2.2and
1804H

the column studies presented in Section 8.1. As d decreases, the beam failure mode
1805H

transitions from a distortional buckling failure to yielding and collapse of the net
section.Figure8.50highlightsthistransitionfortheSSMA550S16254beamconsidered
1806H

382

in this study by comparing the deformed shape at ultimate limit state as hole size
increases.
Elastic buckling controlled failure

Inet/Ig

Yielding and collapse of net section

1.0

0.95

Mtest25/Myg

0.90

0.89

0.90
0.83

0.87

0.92

0.93

Figure8.50SSMA550S16254structuralstudfailuremodetransitionfromdistortionalbucklingtoyielding
atthenetsection

TheobservationsfromthisstudyareusedtoformulateamodifiedDSMdistortional

curve for beams with holes which captures the failure mechanism transition from
yielding at the net crosssection to a distortional type failure mode and limits the
strengthofthebeamtotheyieldmomentatthenetsection:
Distortional Buckling
The nominal flexural strength, Mnd, for distortional buckling shall be calculated in accordance with the
following:
(a) For d d1
Mnd = M ynet

(caponcolumnstrength)

(b) For d1<dd2

ynet
d2
Mnd = M

ynet
d 2 d1
M

(yieldcontroltransition)

(c) For d > d2

Mnd = 1 0.22 M crd

M
y

where
d

d1
d2
d2

0. 5

M
crd
M
y

0 .5

My

(existingDSMdistortionalcurve)

M y M crd

= 0.673( M ynet M y )

= 0.673 1.7 (M ynet M y )1.7 0.7


= (1 0.22(1 d 2 ))(1 d 2 )M y

Mynet = SfnetFy0.80My
(limitreductionofthenetsectionto0.8My)
= Section modulus at the hole(s) referenced to the extreme fiber at first yield

Sfnet

Mcrd

= Critical elastic distortional beam buckling load including hole(s)

383

ThemodifiedDSMdistortionalcurveisaddedinFigure8.51btoFigure8.53basInet/Ig
1807H

180H

decreases, simulating the transition from the existing DSM curve to the net section
strength limit exhibited by the simulated test data. The linear portion of themodified
prediction curve represents the unstiffened strip distortional collapse mechanism and
thenonlinearportionrepresentsacollapsemechanismdrivenbydistortionalbuckling.
ThisproposedmodificationtotheDSMdistortionalpredictioncurvewillbecompared
against the beam experiments database developed in Section 4.3.1 as a part of several
1809H

proposedDSMoptionsconsideredlaterinthischapter.
1.4

1.4
DSM (no hole)
FE 25% CDF imperfections
FE 75% CDF imperfections

0.8

0.8

0.6

0.6

0.4

0.4

0.2

0.2

0.5

1.5

2.5

DSM (no hole)


DSM (proposed, with holes)
FE 25% CDF imperfections
FE 75% CDF imperfections

1.2

Mtest/My

Mtest/My

1.2

distortional slenderness, d=(My /Mcrd)0.5

0.5

1.5

2.5

distortional slenderness, d=(My /Mcrd)0.5

Figure8.51Comparisonofsimulatedbeamstrengths(Inet/Ig=1.0)to(a)theexistingDSMdistortional
bucklingdesigncurveandto(b)theproposedDSMdistortionalbucklingcurveforbeamswithholes
1.4

1.4
DSM (no hole)
FE 25% CDF imperfections
FE 75% CDF imperfections

0.8

0.8

0.6

0.6

0.4

0.4

0.2

0.2

0.5

1.5

distortional slenderness, d=(My /Mcrd)0.5

2.5

DSM (no hole)


DSM (proposed, with holes)
FE 25% CDF imperfections
FE 75% CDF imperfections

1.2

Mtest/My

Mtest/My

1.2

0.5

1.5

2.5

distortional slenderness, d=(My /Mcrd)0.5

Figure8.52Comparisonofsimulatedbeamstrengths(Inet/Ig=0.95)to(a)theexistingDSMdistortional
bucklingdesigncurveandto(b)theproposedDSMdistortionalbucklingcurveforbeamswithholes

384

1.4

1.4
DSM (no hole)
FE 25% CDF imperfections
FE 75% CDF imperfections

0.8

0.8

0.6

0.6

0.4

0.4

0.2

0.2

0.5

1.5

2.5

distortional slenderness, d=(My /Mcrd)0.5

DSM (no hole)


DSM (proposed, with holes)
FE 25% CDF imperfections
FE 75% CDF imperfections

1.2

Mtest/My

Mtest/My

1.2

0.5

1.5

2.5

distortional slenderness, d=(My /Mcrd)0.5

Figure8.53Comparisonofsimulatedbeamstrengths(Inet/Ig=0.90)to(a)theexistingDSMdistortional
bucklingdesigncurveandto(b)theproposedDSMdistortionalbucklingcurveforbeamswithholes

2.4 Presentation and evaluation of DSM options


14B

SixoptionsforextendingDSMtolaterallybracedbeamswithholesareevaluatedin

this section. The options range from simple substitutions in the existing code to more
involved modifications, including the incorporation of the design curve transitions
discussedinSection8.2.2andSection8.2.3forlocalanddistortionalbuckling.
180H

18H

385

2.4.1 DescriptionofDSMoptions
183B

Option 1: Include hole(s) in Mcr determinations, ignore hole otherwise


This method, in presentation, appears identical to currently available DSM expressions
Lateral-Torsional Buckling
The nominal flexural strength, Mne, for lateral-torsional buckling shall be calculated in accordance with the
following:
(a) for Mcre < 0.56 My

Mne = Mcre

(b) for 2.78My Mcre 0.56My M = 10 M 1 10M y


ne
y

(c) for Mcre >2.78My

36M cre

Mne= My

where
Mcre= Critical elastic global beam buckling load (including hole(s))
Local Buckling
The nominal flexural strength, Mnl, for local buckling shall be calculated in accordance with the following:
(a) For l 0.776
Mnl = Mne

(b) For l > 0.776

Mnl = 1 0.15 M crl

where
l

0.4

M ne

M
crl
M ne

0.4

M ne

= M ne M crl

Mcrl = Critical elastic local beam buckling load including hole(s)


Mne=

defined in section above

Distortional Buckling
The nominal flexural strength, Mnd, for distortional buckling shall be calculated in accordance with the
following:
(a) For d 0.673
Mnd = M y

(b) For d >0.673

M crd
Mnd =

1 0.22 M
y

where
d
Mcrd

0. 5

M
crd
M
y

0 .5

My

M y M crd

= Critical elastic distortional beam buckling load including hole(s)

386

Option 2: Include hole(s) in Mcr determinations, Use Mynet everywhere


The only change in this method is to replace My with Mynet
Lateral-Torsional Buckling
The nominal flexural strength, Mne, for lateral-torsional buckling shall be calculated in accordance with the
following:
(a) for Mcre < 0.56 Mynet

Mne = Mcre

(b) for 2.78Mynet Mcre 0.56Mynet M = 10 M 1 10M ynet


ne
ynet

(c) for Mcre >2.78Mynet

36M cre

Mne= Mynet

where
Mcre= Critical elastic global beam buckling load (including hole(s))
Mynet = SfnetFy0.80My
Sfnet = Section modulus at the hole(s) referenced to the extreme fiber at first yield

Local Buckling
The nominal flexural strength, Mnl, for local buckling shall be calculated in accordance with the following:
(a) For l 0.776
Mnl = Mne

(b) For l > 0.776

Mnl = 1 0.15 M crl

where
l

M ne

0.4

M
crl
M ne

0.4

M ne

= M ne M crl

Mcrl = Critical elastic local beam buckling load including hole(s)


Mne = defined in section above
Distortional Buckling
The nominal flexural strength, Mnd, for distortional buckling shall be calculated in accordance with the
following:
(a) For dnet 0.673
Mnd = M ynet

(b) For dnet >0.673


0 .5

M crd M crd
Mnd =


1
0
.
22

M M
ynet ynet

where
dnet
= M ynet M crd

Mcrd

0. 5

M ynet

= Critical elastic distortional beam buckling load including hole(s)

387

Option 3: Cap Mnl and Mnd, otherwise no strength change, include hole(s) in Mcr
This method puts bounds in place and assumes local-global interaction happens at full Mne
Lateral-Torsional Buckling
The nominal flexural strength, Mne, for lateral-torsional buckling shall be calculated in accordance with the
following:
(a) for Mcre < 0.56 My

Mne = Mcre

(b) for 2.78My Mcre 0.56My M = 10 M 1 10M y


ne
y

(c) for Mcre >2.78My

36M cre

Mne= My

where
Mcre= Critical elastic global beam buckling load (including hole(s))
Local Buckling
The nominal flexural strength, Mnl, for local buckling shall be calculated in accordance with the following:
(a) For l 0.776
Mnl = Mne Mynet
(b) For l > 0.776

Mnl = 1 0.15 M crl

where
l

M ne

0.4

M
crl
M ne

0.4

M ne

= M ne M crl

Mcrl = Critical elastic local beam buckling load including hole(s)


Mne=

defined in section above

Mynet = SfnetFy0.80My
Sfnet = Section modulus at the hole(s) referenced to the extreme fiber at first yield

Distortional Buckling
The nominal flexural strength, Mnd, for distortional buckling shall be calculated in accordance with the
following:
(a) For d 0.673
Mnd = My Mynet

(b) For d >0.673


0. 5

M crd M crd
Mnd =

1 0.22 M M
y

where
d
= M y M crd

Mcrd

0 .5

My

= Critical elastic distortional beam buckling load including hole(s)

388

Option 4: Cap Mnl, transition Mnd, include hole(s) in Mcr determinations


This method puts bounds and transition in place, assumes local-global interaction at full Mne
Lateral-Torsional Buckling
The nominal flexural strength, Mne, for lateral-torsional buckling shall be calculated in accordance with the
following:
(a) for Mcre < 0.56 My

Mne = Mcre

(b) for 2.78My Mcre 0.56My M = 10 M 1 10M y


ne
y

(c) for Mcre >2.78My

36M cre

Mne= My

where
Mcre= Critical elastic global beam buckling load (including hole(s))
Local Buckling
The nominal flexural strength, Mnl, for local buckling shall be calculated in accordance with the following:
(a) For l 0.776
Mnl = Mne Mynet

(b) For l > 0.776

Mnl = 1 0.15 M crl


M
ne

where
l

0.4

M
crl
M ne

0.4

M ne

= M ne M crl

Mcrl = Critical elastic local beam buckling load including hole(s)


Mne=

defined in section above

Mynet = SfnetFy0.80My
Sfnet = Section modulus at the hole(s) referenced to the extreme fiber at first yield
Distortional Buckling
The nominal flexural strength, Mnd, for distortional buckling shall be calculated in with the following:
(a) For d d1
Mnd = M ynet

(caponcolumnstrength)

(b) For d1<dd2

ynet
d2
Mnd = M

ynet

d1
d2
M

(yieldcontroltransition)

(c) For d > d2


0. 5

Mnd = 1 0.22 M crd M crd

M M
y y

where
d
= M y M crd

d1
d2
d2

0 .5

My

(existingDSMdistortionalcurve)

= 0.673( M ynet M y )

= 0.673 1.7 (M ynet M y )1.7 0.7


= (1 0.22(1 d 2 ))(1 d 2 )M y

Mynet = SfnetFy0.80My
(limitreductionofthenetsectionto0.8My)
Sfnet
= Section modulus at the hole(s) referenced to the extreme fiber at first yield
Mcrd

= Critical elastic distortional beam buckling load including hole(s)

389

Option 5: Transition Mnl (Option A), transition Mnd, include hole(s) in Mcr determinations
This method puts bounds and transition in place, assumes local-global interaction at full Mne
Lateral-Torsional Buckling
The nominal flexural strength, Mne, for lateral-torsional buckling shall be calculated in accordance with the
following:
(a) for Mcre < 0.56 My

Mne = Mcre

(b) for 2.78My Mcre 0.56My M = 10 M 1 10M y


ne
y

(c) for Mcre >2.78My

36M cre

Mne= My

where
Mcre= Critical elastic global beam buckling load (including hole(s))
Local Buckling
The nominal flexural strength, Mnl, for local buckling shall be calculated in accordance with the following:
(a) For l l1

Mnl = Mne Mynet

(caponbeamstrength)

(b) For l1<ll2

l 2 l l1
Mnl = M

ynet M ynet M l 2

l l 2 l1

(c) For l > l2

Mnl = 1 0.15 M crl

where
l

l1

M ne

0.4

M
crl
M ne

(nonlinearyieldtransitionwhenMynet/Mne1)

0.4

M ne

(DSMlocalbucklingcurve,unchanged)

= M ne M crl
= 0.776( M ynet M ne ) 0.776

l2 = 0.776 2.4(M ynet M ne )3.5 1.4 , Mynet/Mne1


= 0.776, Mynet/Mne>1

l2 = 1 0.15(1 l 2 )

0.8

)(1

l2

0.8

(notransitionwhenMynet/Mne>1)

M ne

Mynet = SfnetFy0.80My
(limitreductionofthenetsectionto0.8My)
Sfnet = Section modulus at the hole(s) referenced to the extreme fiber at first yield
Mcrl = Critical elastic local beam buckling load including hole(s)

Distortional Buckling

SameasOption4

390

Option 6: Transition Mnl (Option B), transition Mnd, include hole(s) in Mcr determinations
This method puts bounds and transition in place, assumes local-global interaction at full Mne
Flexural, Torsional, or Torsional-Flexural Buckling
The nominal axial strength, Pne, for flexural, or torsional- flexural buckling is
2
for c 1.5 Pne = 0.658 c Py

0.877
for c > 1.5 Pne = 2 Py = 0.877Pcre

c
where

Py Pcre

Pcre= Critical elastic global column buckling load (including hole(s))


Py

= AgFy
= gross area of the column

Ag

Local Buckling
The nominal axial strength, Mnl, for local buckling shall be calculated in accordance with the following:
(a) For l l1

Mnl = Mynet (Mne /My)

(caponcolumnstrength)

(b) For l1<ll2

Mnl = M M ne M M ne M l 2 l l1
ynet
ynet
l2

M
y

M
y

(c) For l > l2

Mnl = 1 0.15 M crl

where
l

l1

M
ne

0.4

M
crl
M ne

l l 2 l1

0.4

M ne

(DSMlocalbucklingcurve,unchanged)

= M ne M crl
= 0.776( M ynet M y ) 0.776

l2 = 0.776 2.4(M ynet M y )3.5 1.4

(nonlinearyieldtransition)

l2 = 1 0.15(1 l 2 ) (1 l 2 ) M ne
Mynet = SfnetFy0.80My
0.8

0.8

(limitreductionofthenetsectionto0.8My)

Sfnet = Section modulus at the hole(s) referenced to the extreme fiber at first yield
Mcrl = Critical elastic local beam buckling load including hole(s)
Distortional Buckling

SameasOption4

391

2.5 DSM comparison to beam test simulation database


15B

The six DSM prediction options for coldformed steel beams with holes are

evaluated with the simulated laterally braced beam experiment database developed in
Section 8.2.1andsummarizedin AppendixL.Thesimulateddataiscomparedagainst
182H

183H

DSMpredictionswhileevaluatingdatatrendsagainstmemberslendernessandholesize
(Inet/Ig),andspantodepthratio(L/H).

Figure8.51and Figure8.52comparethesimulatedtestdatatopredictionsforlocal
184H

185H

anddistortionalbucklingcontrolledbeamfailures.Option1isidenticaltotheexisting
DSM approach for beams without holes, except the critical elastic buckling loads (Mcrl,
Mcrd,andMcre)aredeterminedwiththeinfluenceofholes.Option1isobservedtobea
accurate predictor of local buckling controlled failure strengths, although distortional
predictionsareconservativewhendishighandunconservativebyasmuchas20%asd
decreases below 1.5 (see Figure 8.52). The unconservative predictions occur because
186H

Option1doesnotaccountforthecolumnstrengthlimitMynet,nordoesitaccountfora
transitionfromanelasticbucklingcontrolledfailuretoayieldcontrolledfailureat the
netsection.

Option2isobservedtobeaconservativepredictorinFigure8.51andFigure8.52for
187H

18H

high l, and d and demonstrates improved accuracy over Option 1 when slenderness
decreasesandholesizeincreases(see Figure8.52).Option2replacesMyneteverywhere
189H

within the existing DSM formulation, which has the effect of increasing l and d and
decreasingpredictedstrength.Option3testtopredictedtrendsaresimilartoOption1
with increasingly unconservative predictions as slenderness decreases, demonstrating

392

that the Mynet limits on Mnl and Mnd in Option 3 are not fully effective at capturing the
yield transition to the net section. Option 4 is identical to Option 3 except the yield
transition on the DSM distortional curve is employed to provide a more accurate
predictionofthenetsectionyieldinginfluence.Option4demonstratesanimprovement
in distortional bucklingcontrolled prediction accuracy when d < 1, although the
strengthof11beamsareunderpredictedbyupto15%whend=1.3.Option5accurately
predictsthestrengthofthese11beamswiththeaddedtransitiononthelocalbuckling
design curve. (Option 6 is the same as Option 5 because the beams considered are
laterallybraced).

Table 8.1 summarizes the testtopredicted ratio statistics and strength reduction
1820H

factorforthesixDSMoptions(seeEq.(8.1)foradefinitionof).Nooneoptionstands
182H

outabovetherestwhenstudyingthetable,althoughtheobservationsfrom Figure8.51
182H

and Figure 8.52 support Options 3,4, and 5(6) as the methods most closely tied to
1823H

underlying collapse mechanisms at ultimate limit state. The observations from this
comparisonwillbeemployedalongwiththeDSMcomparisontothebeamexperimental
databaseinthenextsectiontosupporttherecommendedDSMmodifications.
Table8.4DSMtesttopredictedstatisticsforlaterallybracedbeamsimulations
Option

Description

Local buckling

SD
0.09
0.89

Distortional buckling

SD
0.13
0.86

My everywhere

Mean
1.07

Mynet everywhere

1.05

0.10

0.88

50

1.07

0.12

0.86

154

Cap Mnl, Mnd

1.07

0.09

0.89

44

1.06

0.13

0.86

160

Transition Mnd, Cap Mnl

1.07

0.09

0.89

44

1.06

0.13

0.86

160

Transition Mnd and Mnl (Option A, B)

1.01

0.11

0.87

72

1.09

0.12

0.87

132

5,6

393

# of tests
44

Mean
1.06

# of tests
160

1.5

1.5
Option 1 - My everywhere

Option 2 - Mynet everywhere


1
Mn/My

Mn/My

0.5

0.5

1.5

2.5

0.5

0.5

1.5

0.5

Option 4 - cap Mn , transition Mnd

1
Mn/My

Mn/My

DSM Local Prediction


Local Controlled Test

0.5

0.5

1.5

2.5

0.5

0.5

1.5

0.5

2.5

2.5

0.5

l=(My /Mcrl)

l=(My /Mcrl)

1.5

1.5
Option 5 - transition Mn (Option A), transition Mnd

Option 6 - transition Mn (Option B), transition Mnd

1
Mn/My

1
Mn/My

1.5
Option 3 - cap Mn , Mnd

0.5

2.5

l=(My /Mcrl)

1.5

2
0.5

l=(My /Mcrl)

0.5

1.5

2.5

0.5

0.5

0.5

1.5

2
0.5

l=(My /Mcrl)

l=(My /Mcrl)

Figure8.54Comparisonofsimulatedteststrengthstopredictionsforlaterallybracedbeamswithlocalbucklingcontrolledfailuresasafunctionoflocal
slenderness

394

1.5

1.5
Option 1 - My everywhere

Option 2 - Mynet everywhere


1
Mn/My

Mn/My

0.5

0.5

1.5

2.5

0.5

0.5

0.5

2.5

2.5

Option 4 - cap Mn , transition Mnd

1
Mn/My

1
Mn/My

2.5

1.5
Option 3 - cap Mn , Mnd

0.5

0.5

1.5

2.5

0.5

0.5

0.5

1.5
0.5

d=(My /Mcrd)

d=(My /Mcrd)

1.5

1.5
Option 5 - transition Mn (Option A), transition Mnd

Option 6 - transition Mn (Option B), transition Mnd

1
Mn/My

1
Mn/My

dnet=(Mynet/Mcrd)

1.5

0.5

1.5
0.5

d=(My /Mcrd)

DSM Dist. Prediction


Dist. Controlled Test

0.5

1.5

2.5

0.5

0.5

0.5

1.5

2
0.5

d=(My /Mcrd)

d=(My /Mcrd)

Figure8.55Comparisonofsimulatedteststrengthstopredictionsforlaterallybracedbeamswithdistortionalbucklingcontrolledfailuresasafunctionof
distortionalslenderness

395

1.5

Mtest/Mn

Mtest/Mn

1.5

0.5

0.5

Option 1 - My everywhere
0

0.5

1.5

2.5

0.5

1.5

0.5

Mtest/Mn

Mtest/Mn

1.5

0.5

1.5

2.5

0.5

Mtest/Mn

Mtest/Mn

2.5

0.5

1.5

0.5

0.5

Option 5 - transition Mn (Option A), transition Mnd

Option 6 - transition Mn (Option B), transition Mnd

1.5

1.5

l=(My /Mcrl)

1.5

0.5

2.5

l=(My /Mcrl)

Option 4 - cap Mn , transition Mnd

0.5

0.5

Option 3 - cap Mn , Mnd


l
0.5

2.5

l=(My /Mcrl)

1.5

2
0.5

l=(My /Mcrl)

Local Controlled

Option 2 - Mynet everywhere

2.5

0.5

0.5

1.5

2
0.5

l=(My /Mcrl)

l=(My /Mcrl)

Figure8.56Testtopredictedratiosforlocalbucklingcontrolledsimulatedlaterallybracedbeamfailuresasafunctionoflocalslenderness

396

1.5

Mtest/Mn

Mtest/Mn

1.5

0.5

Distortional Controlled
0.5

Option 1 - My everywhere
0

0.5

Option 2 - Mynet everywhere


1.5

2.5

0.5

0.5

Mtest/Mn

Mtest/Mn

1.5

0.5

0.5

1.5

2.5

0.5

Mtest/Mn

Mtest/Mn

2.5

2.5

0.5

1.5

0.5

0.5

Option 5 - transition Mn (Option A), transition Mnd

Option 6 - transition Mn (Option B), transition Mnd

1.5

1.5

d=(My /Mcrd)

1.5

0.5

d=(My /Mcrd)

Option 4 - cap Mn , transition Mnd

0.5

2.5

0.5

Option 3 - cap Mn , Mnd


l
0

dnet=(Mynet/Mcrd)

1.5

1.5
0.5

d=(My /Mcrd)

2.5

0.5

0.5

1.5

2
0.5

d=(My /Mcrd)

d=(My /Mcrd)

Figure8.57Testtopredictedratiosfordistortionalbucklingcontrolledsimulatedlaterallybracedbeamfailuresasafunctionofdistortionalslenderness

397

1.5

Mtest/Mn

Mtest/Mn

1.5

0.5

0.5

Option 1 - My everywhere
0
0.5

0.6

0.7

0.8

0.9

0
0.5

0.6

0.7

1.5

1.5

0.5

0.7

0.8

0.9

0.6

0.7

0.9

0.9

Inet/Ig
1.5

Mtest/Mn

Mtest/Mn

0.8

0
0.5

1.5

0.5

0.5

Option 5 - transition Mn (Option A), transition Mnd

Option 6 - transition Mn (Option B), transition Mnd

0.6

Option 4 - cap Mn , transition Mnd

Inet/Ig

0
0.5

0.9

0.5

Option 3 - cap Mn , Mnd


l
0.6

0.8
Inet/Ig

Mtest/Mn

Mtest/Mn

Inet/Ig

0
0.5

Local Controlled

Option 2 - Mynet everywhere

0.7

0.8

0.9

0
0.5

Inet/Ig

0.6

0.7

0.8
Inet/Ig

Figure8.58Testtopredictedratiosforsimulatedlocalbucklingcontrolledlaterallybracedbeamfailuresasafunctionofnetcrosssectionalmomentofinertiato
grosscrosssectionalmomentofinertia

398

1.5

Mtest/Mn

Mtest/Mn

1.5

0.5

0.5

Option 1 - My everywhere
0
0.5

0.6

0.7

0.8

0.9

0
0.5

0.6

0.7

1.5

1.5

0.5

0.7

0.8

0.9

0.6

0.7

0.9

0.9

Inet/Ig
1.5

Mtest/Mn

Mtest/Mn

0.8

0
0.5

1.5

0.5

0.5

Option 5 - transition Mn (Option A), transition Mnd

Option 6 - transition Mn (Option B), transition Mnd

0.6

Option 4 - cap Mn , transition Mnd

Inet/Ig

0
0.5

0.9

0.5

Option 3 - cap Mn , Mnd


l
0.6

0.8
Inet/Ig

Mtest/Mn

Mtest/Mn

Inet/Ig

0
0.5

Distortional Controlled

Option 2 - Mynet everywhere

0.7

0.8

0.9

0
0.5

Inet/Ig

0.6

0.7

0.8
Inet/Ig

Figure8.59Testtopredictedratiosforsimulateddistortionalbucklingcontrolledlaterallybracedbeamfailuresasafunctionofnetcrosssectionalmomentof
inertiatogrosscrosssectionalmomentofinertia

399

1.5

Mtest/Mn

Mtest/Mn

1.5

0.5

0.5

Option 1 - My everywhere
0

10

1.5

1.5

0.5

10

Mtest/Mn

Mtest/Mn

10

10

L/H
1.5

0.5

0.5

Option 5 - transition Mn (Option A), transition Mnd

Option 6 - transition Mn (Option B), transition Mnd

1.5

10

Option 4 - cap Mn , transition Mnd

L/H

0.5

Option 3 - cap Mn , Mnd


l
0

6
L/H

Mtest/Mn

Mtest/Mn

L/H

Local Controlled

Option 2 - Mynet everywhere

10

L/H

6
L/H

Figure8.60Testtopredictedratiosforsimulatedlocalbucklingcontrolledlaterallybracedbeamfailuresasafunctionofcolumnlength,L,tobeamdepth,H

400

1.5

Mtest/Mn

Mtest/Mn

1.5

0.5

Distortional Controlled
0.5

Option 1 - My everywhere
0

Option 2 - Mynet everywhere

10

1.5

1.5

0.5

10

Mtest/Mn

Mtest/Mn

10

10

L/H
1.5

0.5

0.5

Option 5 - transition Mn (Option A), transition Mnd

Option 6 - transition Mn (Option B), transition Mnd

1.5

10

Option 4 - cap Mn , transition Mnd

L/H

0.5

Option 3 - cap Mn , Mnd


l
0

6
L/H

Mtest/Mn

Mtest/Mn

L/H

10

L/H

6
L/H

Figure8.61Testtopredictedratiosforsimulateddistortionalbucklingcontrolledlaterallybracedbeamfailuresasafunctionofcolumnlength,L,toH

401

2.6 DSM comparison to experimental beam database


16B

The six DSM options are now compared to the laterally braced beam experiment

database first assembled in Section 4.3.1. The database contains the elastic buckling
1824H

propertiesofeachbeam,includingthepresenceofholesandtheinfluenceofboundary
conditions,aswellasthetestedstrengths. Figure8.62through Figure8.65comparethe
1825H

1826H

experiment strengthstoDSMpredictionsforlocalanddistortionalbucklingcontrolled
beamfailures.(Thelocalanddistortionalslendernessisobtainedwiththepurelocal
anddistortionalelasticbucklingloadsLandDinthisstudy,nottheLHandDHmodes
describedinSection4.3).Thetestedstrengthsarelowerthanthepredictionsoverawide
1827H

range of local and distortional slenderness. These trends were first observed in a
preliminary DSM comparison (Moen and Schafer 2007a), and possible reasons for the
difference between test and predictions were hypothesized, including experimental
error,errorinthedeterminationofelasticbucklingloads,andtheinfluenceoftheangle
strapsonthecalculationofthedistortionalcriticalelasticbucklingload.Thebeamsin
thedatabasehaverelativelysmallholes,withInet/Igrangingfrom0.96to0.99asshownin
Figure8.64andFigure8.65,whichsuggeststhatthepresenceofholesshouldnothavea
182H

1829H

significantimpactontestedstrength.Thetesttopredictedstatisticsarethesameforthe
sixDSMoptionsasshownin Table8.5.Itisconcludedthattheexperimentaldatabase,
1830H

initscurrentform,cannotbeusedtoevaluatetheproposedDSMmodifications.Future
workisplannedtoinvestigatethedifferencesbetweentheDSMpredictionsandtested
strengthsforthisdata.Inaddition,morerecenttestsoncoldformedsteelbeamswith

402

holes will be added to the database. Experiments on beams with larger holes are also
needed.
Table8.5DSMtesttopredictedratiostatisticsforbeamexperiments
Option

Description

Local buckling
SD

0.12
0.85

# of tests
55

Mean
0.87

Distortional buckling
SD

0.14
0.81

# of tests
89

My everywhere

Mean
0.88

Mynet everywhere

0.88

0.12

0.85

55

0.87

0.14

0.81

89

Cap Mnl, Mnd

0.88

0.12

0.85

55

0.87

0.14

0.81

89

Transition Mnd, Cap Mnl

0.88

0.12

0.85

55

0.87

0.14

0.81

89

Transition Mnd and Mnl (Option A, B)

0.88

0.12

0.85

55

0.87

0.14

0.81

89

5,6

403

1.5

1.5
Option 1 - My everywhere

Option 2 - Mynet everywhere


1
Mn/My

Mn/My

0.5

0.5

1.5

2.5

0.5

0.5

1.5

0.5

Option 4 - cap Mn , transition Mnd

1
Mn/My

Mn/My

DSM Local Prediction


Local Controlled Test

0.5

0.5

1.5

2.5

0.5

0.5

1.5

0.5

2.5

2.5

0.5

l=(My /Mcrl)

l=(My /Mcrl)

1.5

1.5
Option 5 - transition Mn (Option A), transition Mnd

Option 6 - transition Mn (Option B), transition Mnd

1
Mn/My

1
Mn/My

1.5
Option 3 - cap Mn , Mnd

0.5

2.5

l=(My /Mcrl)

1.5

2
0.5

l=(My /Mcrl)

0.5

1.5

2.5

0.5

0.5

0.5

1.5

2
0.5

l=(My /Mcrl)

l=(My /Mcrl)

Figure8.62Comparisonofexperimentalteststrengthstopredictionsforlaterallybracedbeamswithlocalbucklingcontrolledfailuresasafunctionoflocal
slenderness

404

1.5

1.5
Option 1 - My everywhere

Option 2 - Mynet everywhere


1
Mn/My

Mn/My

0.5

0.5

1.5

2.5

0.5

0.5

0.5

2.5

2.5

Option 4 - cap Mn , transition Mnd

1
Mn/My

1
Mn/My

2.5

1.5
Option 3 - cap Mn , Mnd

0.5

0.5

1.5

2.5

0.5

0.5

0.5

1.5
0.5

d=(My /Mcrd)

d=(My /Mcrd)

1.5

1.5
Option 5 - transition Mn (Option A), transition Mnd

Option 6 - transition Mn (Option B), transition Mnd

1
Mn/My

1
Mn/My

dnet=(Mynet/Mcrd)

1.5

0.5

1.5
0.5

d=(My /Mcrd)

DSM Dist. Prediction


Dist. Controlled Test

0.5

1.5

2.5

0.5

0.5

0.5

1.5

2
0.5

d=(My /Mcrd)

d=(My /Mcrd)

Figure8.63Comparisonofexperimentalteststrengthstopredictionsforlaterallybracedbeamswithdistortionalbucklingcontrolledfailuresasafunctionof
distortionalslenderness

405

1.5

Mtest/Mn

Mtest/Mn

1.5

0.5

0.5

Option 1 - My everywhere
0
0.5

0.6

0.7

0.8

0.9

0
0.5

0.6

0.7

1.5

1.5

0.5

0.7

0.8

0.9

0.6

0.7

0.9

0.9

Inet/Ig
1.5

Mtest/Mn

Mtest/Mn

0.8

0
0.5

1.5

0.5

0.5

Option 5 - transition Mn (Option A), transition Mnd

Option 6 - transition Mn (Option B), transition Mnd

0.6

Option 4 - cap Mn , transition Mnd

Inet/Ig

0
0.5

0.9

0.5

Option 3 - cap Mn , Mnd


l
0.6

0.8
Inet/Ig

Mtest/Mn

Mtest/Mn

Inet/Ig

0
0.5

Local Controlled

Option 2 - Mynet everywhere

0.7

0.8

0.9

0
0.5

Inet/Ig

0.6

0.7

0.8
Inet/Ig

Figure8.64Testtopredictedratiosforexperimentallocalbucklingcontrolledlaterallybracedbeamfailuresasafunctionofnetcrosssectionalmomentofinertia
togrosscrosssectionalmomentofinertia

406

1.5

Mtest/Mn

Mtest/Mn

1.5

0.5

0.5

Option 1 - My everywhere
0
0.5

0.6

0.7

0.8

0.9

0
0.5

0.6

0.7

1.5

1.5

0.5

0.7

0.8

0.9

0.6

0.7

0.9

0.9

Inet/Ig
1.5

Mtest/Mn

Mtest/Mn

0.8

0
0.5

1.5

0.5

0.5

Option 5 - transition Mn (Option A), transition Mnd

Option 6 - transition Mn (Option B), transition Mnd

0.6

Option 4 - cap Mn , transition Mnd

Inet/Ig

0
0.5

0.9

0.5

Option 3 - cap Mn , Mnd


l
0.6

0.8
Inet/Ig

Mtest/Mn

Mtest/Mn

Inet/Ig

0
0.5

Distortional Controlled

Option 2 - Mynet everywhere

0.7

0.8

0.9

0
0.5

Inet/Ig

0.6

0.7

0.8
Inet/Ig

Figure8.65Testtopredictedratiosforexperimentaldistortionalbucklingcontrolledlaterallybracedbeamfailuresasafunctionofnetcrosssectionalmomentof
inertiatogrosscrosssectionalmomentofinertia

407

2.7 Recommendations DSM for beams with holes


17B

Options3,4,5,and6arepresentedasviableproposalsforextendingDSMtobeams

with holes. This recommendation is based on the testtopredicted statistics and data
trends from the simulation studies presented in Section 8.2.5, and also considers the
183H

efforttoimplementthemodificationsandtheireaseofusebydesignengineers.Option
3accountsforthereductioninbeamstrengthfromthepresenceofholesbylimitingMnl
and Mnd to Mynet. This is a simple modification to implement in the Specification and
avoids additional calculation work for a design engineer (except for that required to
calculatethecriticalelasticbucklingloadsincludingtheinfluenceoftheholes).Options
4and5arerefinementsofOption3,wherethecaponMnlandMndbecomesatransition
from an elastic buckling controlled failure mode to a yield controlled failure at Mynet.
These two methods require additional effort from the designer when compared to
Option3,buttheyhaveanimportantadvantage.Options4,5,and6aremoreclosely
tied to the failure mechanisms influencing column strength because they capture the
yield transition to the net section in their predictions. The transitions increase the
probability that strength will be accurately predicted for general beam and hole
geometries.Option5hastheadditionaladvantageofcapturingtheinfluenceofayield
transition for closed crosssections that do not experience distortional buckling.
Additionalnonlinearfiniteelementsimulationsandexperimentsareneededtovalidate
the proposed modifications to theDirect Strength Method for beamssubject to lateral
torsionalbucklingatultimatelimitstate.

408

Chapter 9
Conclusions and proposed future work
8B

9.1 Conclusions
50B

ProposedDirectStrengthMethoddesignequationsarenowinplaceforcoldformed

steelmemberswithholes.Thedevelopmentofthemethodwasinitiatedwiththinshell
finite element eigenbuckling studies in ABAQUS on thin plates and full coldformed
steelmemberswithholes.Thebucklingoftheunstiffenedstripsadjacenttoaholeina
thin plate influenced, and sometimes controlled, the critical elastic buckling stress of
individualcrosssectionelements.Unstiffenedstripbucklingwasalsocloselyassociated
withdistortionalbucklingmodesatthelocationoftheholesinCsectioncolumnsand
beams. Large holes and closelyspaced holes locally stiffened thin rectangular plates
and the webs of Csection columns, resulting in buckling away from the holes. The
elasticbucklingstudiesledtousefuldesignguidelinesandtools,includingholespacing
limits (which prevent cumulative reductions in elastic stiffness along the length of a

409

member)andsimplifiedelasticbucklingpredictionmethodsforlocal,distortional,and
globalbucklingdevelopedasanalternativetofiniteelementeigenbucklinganalysis.

TheviabilityoftheDSMframeworkforcoldformedsteelmemberswithholeswas

established early in this research using existing test results and the elastic buckling
properties of coldformed steel column and beam specimens with holes. Additional
experimental work evaluated the influence of holes on the loaddeformation response
and failure mechanisms for short and intermediate length Csection columns. During
the experiments, holes were observed to locally stiffen the web of the intermediate
lengthCsectioncolumnsandpreventeddynamicmodeswitching(fromlocalbuckling
to distortional buckling) near peak load. Holes were also observed to decrease post
peakductilityforcolumnswhentheholesizewaslargerelativetothewebwidth(e.g.,
the362S16233specimens).

Results from the experimental program were used to validate a nonlinear finite

elementmodelingprotocol.Aconcertedeffortwasmadetosimulatetheinitialstateofa
coldformedsteelmemberintheprotocol,includingimperfectionmagnitudesbasedon
measurement statistics and residual stresses and initial plastic strains from the cold
forming process predicted with a mechanicsbased approach. The nonlinear finite
elementmodelingcapabilitywasusedtoconstructalargedatabaseofsimulatedcolumn
and beam experiments with a wide range of hole sizes, spacings, and Csection
dimensions.Simulationresultsdemonstratedthatascrosssectiondistortionalor local
slendernessdecreased,thefailureofacoldformedsteelmemberwithholesoccurredby
yielding and collapse of the unstiffened stripsat thenet crosssection. Collapse of the

410

unstiffened strips sometimes triggered unstable global failure modes in columns with
large holes, i.e., as hole size approached Anet=0.60Ay. (Global instabilities caused by
yieldingatpeakloadwerenotstudiedforbeamswithholesinthisthesis,onlylaterally
bracedbeamswereconsidered.)ModificationstothelocalanddistortionalDSMcurves
were made to account for this unique netsection failure mechanism with a deliberate
transitionandcaponmemberstrength.ThefinalproposedDSMmethodformembers
withholeswasvalidatedwithexistingexperimentaldataandthesimulatedexperiments
database.

9.2 Futurework
51B

Severalinterestingfutureresearchtopicsresultedfromtheelasticbucklingstudies,

experiments,andnonlinearfiniteelementsimulationsinthisthesis.Futureresearchis
plannedtofollowuponmanyoftheseideasandquestions.Themajorpointsoffuture
study,organizedbyresearchtopic,arelistedbelow.

ThinshellfiniteelementmodelinginABAQUS(Chapter2)
The S9R5 meshing guidelines developed in this thesis were developed primary for
eigenbucklinganalyses.Meshingguidelineswhichensureaccurateresultsinnonlinear
finite element simulations are also needed. Studies are ongoing to develop rules for
determining the minimum number of throughthickness finite element integration
points, themesh density required for linear and quadratic finite element formulations,
andlimitsoninitialelementdistortionandcurvature.

411

Elasticbucklingofcoldformedsteelcrosssectionalelementswithholes(Chapter3)
1.

2.

3.
4.

The simplified elastic buckling prediction method presented in this thesis for
unstiffened elements loaded with uniaxial compression is empirically derived.
A mechanicsbased unstiffened element prediction method is warranted as a
topicoffutureresearchtoimprovethegeneralityofthemethod.
Anelementbasedelasticbucklingpredictionmethodwhichaccountsforstress
gradientsonunstiffenedelementswithholesisneededtoaddressadesigncase
engineersmayencounterinpractice.
Elastic buckling studies are planned to develop elementbased simplified
methodsforholepatternsfoundinstorageracks.
The elementbased elastic buckling prediction methods provide a convenient
method to calculate Fcr (the critical elastic buckling stress) for general hole
shapes,sizes,andspacingsforuseintheAISIS10007effectivewidthmethod.
Work is planned to evaluate introduce these simplified approaches into the
effectivewidthmethod.

Elasticbucklingofcoldformedsteelmemberswithholes(Chapter4)
1. Yu and Davis, OrtizColberg, Rhodes and MacDonald, Rhodes and Schnieder,
andPuetal.performedtestsoncolumnspecimenswithmultiplediscreteholes
or hole patterns. The elastic buckling properties and tested strengths of these
specimenswillbeaddedtotheexperimentdatabase,inadditiontotestsonrack
sections.
2. Automated elastic buckling modal identification tools are needed to identify
local, distortional, and global buckling modes in thinshell finite element
eigenbuckling analysis. Research is ongoing to develop this capability with an
implementationsimilartothatoftheconstrainedfinitestripmethod.
3. Work continues on the development and validation of the CUFSM elastic
bucklingapproximatemethodsdevelopedandtheextensionofthesemethodsto
members with hole patterns (e.g., storage racks). A general procedure for
implementing CUFSM constraints in the local buckling prediction method is
needed.Also,thecurrentassumptionthatthewarpingtorsionconstantCw=0ata
hole produces conservative global elastic buckling predictions for columns and
beams. Additional research is needed to derive a mechanicsbased
approximationforCwatahole.

Experimentsoncoldformedsteelcolumnswithholes(Chapter5)
1.

Amoredefinitivemethodofmeasuringthebasemetalthicknessofcoldformed
steelmemberswithazincgalvaniccoatingisneeded.Currentstandardpractice
istoremovethezinccoatingwithhydrochloricacidoraferricchloridesolution.
It is difficult to know when all of the zinc has been removed though since the

412

2.

zinc chemically interacts with the base metal during the initial application.
Experiments are planned to determine the influence of the zinc coating on
ultimatestrength.
Research work is planned to evaluate the influence of sheet coiling on the
measuredyieldstressintensilecoupons.IthasbeenhypothesizedbyProfessor
Rasmussen at the University of Sydney that the same coiling curvature which
causes residual stresses in coldformed steel members also affects yield stress
measurementsintensiletests.

Residualstressesandplasticstrainsincoldformedsteelmembers(Chapter6)
1. Experimental work is planned to validate the prediction model presented in
Chapter6relatingcoilingresidualstressestothecoilingradius,sheetthickness,
andyieldstress.
2. Research is ongoing to evaluate how ABAQUS metal plasticity laws use the
residualstressandinitialplasticstraininformationandtodetermineifkinematic
hardeningoradifferentmixedhardeningruleisrequiredtoaccuratelysimulate
thecoldworkofformingeffectonloaddeformationresponse.
3. Nonlinearfiniteelementstudiesareplannedtoidentifytheinfluenceofthrough
thicknessresidualstressesandplasticstrainsontheloaddeformationresponse,
ultimate strength, and failure mechanisms of coldformed steel beams and
columns.
4. Hancocketal.providesamethodwhichaccountsforthecoldworkofforming
in the corners of coldformed steel crosssections when calculating ultimate
strength(Hancocketal.2001).TheresearchinChapter6providesnewinsight
intotherelationshipbetweenresidualstressesandinitialplasticstrainsfromthe
manufacturing process. Research work is planned to revisit Hancocks cold
work of forming method to determine if it can be supplemented with this new
research.
5. Thecurrentresidualstresspredictionmethodassumesanelasticperfectlyplastic
materialmodel.Researchworkisongoingtointroducetheeffectofsteelstrain
hardeningintothepredictionmethod.

Nonlinearfiniteelementmodelingofcoldformedsteelmembers(Chapter7)
1. Theuseofmeasuredimperfectionmagnitudesinsteadofstatisticaldistributions
is warranted as a topic of future study, especially the use of a flared cross
section,includingflangewebanglesoffof90degrees.
2. InitiatingplasticityinABAQUSatthematerialsproportionallimitreducedthe
predictedstrengthbyupto20%whencomparedtoexperimentsinChapter7.A
study is planned to simulate a single finite element in tension to evaluate the
ABAQUS implementation of metal plasticity and determine the source of the
discrepancy.

413


TheDirectStrengthMethodformemberswithholes(Chapter8)
1.
2.

AdditionalvalidationstudiesareplannedtocomparetheproposedDSMHoles
methodologytotheAISIS10007effectivewidthdesignmethod.
Nonlinear finite element studies of other DSM prequalified crosssections (e.g.,
Zsectionsandhatsections)aswellasracksectionswithholepatternsarealso
plannedtoexpandthesimulationdatabase.

414

References
ABAQUS.(2007a).ABAQUS/StandardVersion6.71.,DassaultSystmes,
http://www.simulia.com/,Providence,RI.

ABAQUS.(2007b).Personalcommunication,S9R5elementsectionnumbers.Dassault
Systmes,http://www.simulia.com/,Providence,RI.

AISIS100.(2007).NorthAmericanSpecificationfortheDesignofColdFormedSteelStructural
Members,AmericanIronandSteelInstitute,Washington,D.C.

AISITS202.(2001).StubColumnTestMethodforEffectiveAreaofColdformedSteel
Columns.2001NorthAmericanColdFormedSteelSpecification,AmericanIronand
SteelInstitute,Washington,D.C.

AISI.(2006).DirectStrengthMethodDesignGuide,AmericanIronandSteelInstitute,
Washington,D.C.

ASTM.(2004).E8M04,StandardTestMethodsforTensionTestingofMetallicMaterials
(Metric).ASTMInternational,WestConshohocken,PA.

Batson,K.D.(1992).FlexuralBehaviorofWebswithOpenings,M.S.Thesis,Universityof
MissouriRolla,Rolla.

Bernard,E.S.(1993).FlexuralBehaviourofColdFormedProfiledSteelDecking,Ph.D.Thesis,
UniversityofSydney,Australia,Sydney.

Chajes,A.(1974).PrinciplesofStructuralStability,PrenticeHallCollegeDiv,EnglewoodCliffs,
NJ.

Chen,W.F.,andHan,D.J.(1988).PlasticityforStructuralEngineers,SpringerVerlag,NewYork,
NY.

Cook,R.D.(1989).ConceptsandApplicationsofFiniteElementAnalysis,J.Wiley&Sons,New
York,NY.

CRC.(2003).StandardMathematicalTablesandFormulae,CRCPress,NewYork,NY.

Crisfield,M.A.(1981).Afastincremental\iterationsolutionprocedurethathandlessnap
through.ComputersandStructures,13,5562.

Dat,D.T.(1980).TheStrengthofColdFormedSteelColumns.CornellUniversityDepartment
ofStructuralEngineeringReportNo.804,Ithaca,NY.

435H

436H

415

DeBatista,E.M.,andRodrigues,F.C.(1992).Residualstressmeasurementsoncoldformed
profiles.ExperimentalTechniques,16(5),2529.

Galambos,T.(1998a).AppendixB.3,TechnicalMemorandumNo.3:StubColumnTest
Procedure.GuidetoStabilityDesignCriteriaforMetalStructures,5thEdition,John
Wiley&Sons,NewYork,NY.

Galambos,T.(1998b).GuidetoStabilityDesignCriteriaforMetalStructures,5thEdition,John
Wiley&Sons,NewYork,NY.

Hancock,G.J.,Murray,T.,andEllifritt,D.(2001).ColdformedsteelstructurestotheAISI
specification,MarcelDekker,Inc.,NewYork,NY.

Hill,R.(1950).TheMathematicalTheoryofPlasticity,OxfordUniversityPress,London,England.

Ingvarsson,L.(1975).Coldformingresidualstresses,effectofbuckling.Proceedings,Third
AnnualSpecialtyConferenceonColdFormedSteelStructures,UniversityofMissouriRolla,
85119.

Key,P.,andHancock,G.J.(1993).ATheoreticalInvestigationoftheColumnBehaviourof
ColdFormedSquareHollowSections.ThinWalledStructures,16,3164.

Kwon,Y.B.(1992).PostBucklingBehaviourofThinWalledSections,Ph.D.Thesis,University
ofSydney,Australia.

Labview.(2005).Labview,Version8.NationalInstruments,www.labview.com,Austin,TX.

Mathworks.(2007).Matlab7.5.0(R2007b).Mathworks,Inc.,www.mathworks.com.

Moen,C.D.,andSchafer,B.W.(2007a).Directstrengthdesignforcoldformedsteelmembers
withperforations,ProgressReport#3.AmericanIronandSteelInstitute,Washington,
D.C.

Moen,C.D.,andSchafer,B.W.(2007b).Directstrengthdesignforcoldformedsteelmembers
withperforations,ProgressReport#4.AmericanIronandSteelInstitute,Washington,
D.C.

Oden,J.T.,andMartins,J.A.C.(1985).Modelsandcomputationalmethodsfordynamicfriction
phenomena.ComputerMethodsinAppliedMechanicsandEngineering,52,527634.

OrtizColberg,R.A.(1981).TheLoadCarryingCapacityofPerforatedColdFormedSteel
Columns,M.S.Thesis,CornellUniversity,Ithaca,NY.

Porter,F.C.(1991).ZincHandbook,MarcelDekker,NewYork,NY.

Powell,G.,andSimons,J.(1981).Improvediterativestrategyfornonlinearstructures.
InternationalJournalforNumericalMethodsinEngineering,17,14551467.

437H

438H

416

Quach,W.M.,Teng,J.G.,andChung,K.F.(2004).Residualstressesinsteelsheetsdueto
coilinganduncoiling:aclosedformanalyticalsolution.EngineeringStructures,26,1249
1259.

Quach,W.M.,Teng,J.G.,andChung,K.F.(2006).Finiteelementpredictionsofresidual
stressesinpressbrakedthinwalledsteelsections.EngineeringStructures,28,16091619.

Ramm,E.(1981).Strategiesfortracingnonlinearresponsenearlimitpoints.NonlinearFinite
ElementAnalysisinStructuralMechanics:ProceedingsoftheEuropeUSWorkshop,6389.

Schafer,B.W.(1997).Coldformedsteelbehavioranddesign:analyticalandnumerical
modelingofelementsandmemberswithlongitudinalstiffeners,Ph.D.Thesis,Cornell
University,Ithaca,NY.

Schafer,B.W.,anddny,S.(2006).Bucklinganalysisofcoldformedsteelmembersusing
CUFSM:conventionalandconstrainedfinitestripmethods.EighteenthInternational
SpecialtyConferenceonColdFormedSteelStructures,Orlando,FL.

Schafer,B.W.,andPekz,T.(1998).Computationalmodelingofcoldformedsteel:
characterizinggeometricimperfectionsandresidualstresses.JournalofConstructional
Research,47,193210.

Schafer,B.W.,Sarawit,A.,andPekz,T.(2006).Complexedgestiffenersforthinwalled
members.JournalofStructuralEngineering,132(2),212226.

Schuster,R.M.(1992).TestingofPerforatedCStudSectionsinBending.Universityof
Waterloo,Ontario,Canada.

Shan,M.Y.,andLaBoube,R.A.(1994).BehaviorofWebElementsWithOpeningsSubjectedto
Bending,Shear,andtheCombinationofBendingandShear.CivilEngineeringStudy942,
ColdformedSteelSeries,UniversityofMissouriRolla,Rolla.

Shanley,F.R.(1957).StrengthofMaterials,McGrawHillBookCompany,NewYork,NY.

Silvestre,N.,andCamotim,D.(2005).Localplateanddistortionalpostbucklingbehaviorof
coldformedsteellippedchannelcolumnswithintermediatestiffeners.Seventeenth
InternationalSpecialtyConferenceonColdFormedSteelStructures:RecentResearchand
DevelopmentsinColdFormedSteelDesignandConstruction,Orlando,FL,UnitedStates,1
18.

SSMA.(2001).ProductTechnicalInformation,ICBOER4943P.SteelStudManufacturers
Association,www.ssma.com.

Timoshenko,S.P.,Gere,JamesM.(1961).TheoryofElasticStability,McGrawHill,NewYork.

USSteel.(1985).TheMaking,Shaping,andTreatingofSteel,10thEdition,Herbick&Held,
Pittsburgh,PA.

439H

417

vonKarman,T.,Sechler,E.E.,andDonnell,L.H.(1932).Thestrengthofthinplatesin
compression.TransactionsASME,54(APM545).

Wang,X.P.,Lam,S.S.E.,andChung,K.F.(2006).Crosssectiondistortionduetocuttingofcold
formedsteellippedCsection.ThinWalledStructures,44,271280.

Weng,C.C.,andPekz,T.(1990).ResidualStressesinColdFormedSteelMembers.ASCE
JournalofStructuralEngineering,116(6),16111625.

WesternStatesClayProductsAssociation.(2004).DesignGuideforAnchoredBrickVeneerover
SteelStudSystems,www.brick_wscpa.org,Seattle,WA.

Winter,G.(1947).Strengthofthinsteelcompressionflanges.CornellUniversityEngineering
ExperimentStationReprintNo.32.

Yu,C.(2005).Distortionalbucklingofcoldformedsteelmembersinbending,Ph.D.Thesis,
JohnsHopkinsUniversity,Baltimore.

Yu,C.,andSchafer,B.W.(2006).Distortionalbucklingtestsoncoldformedsteelbeams.ASCE
JournalofStructuralEngineering,132(4),515528.

Yu,C.,andSchafer,B.W.(2007).EffectofLongitudinalStressGradientsonElasticBucklingof
ThinPlates.ASCEJournalofStructuralEngineering,133(4),452463.

Yu,W.W.(2000).ColdFormedSteelDesign,JohnW.Wiley&Sons,NewYork,NY.

Ziemian,R.,andMcGuire,W.(2005).MASTANversion3.0.4,www.mastan.com.Bucknell
University,Lewisburg,PA.

40H

41H

418

Appendix A
ABAQUS input file generator in Matlab
9B

The finite element models in this thesis were generated with a custom Matlab

program which assembles a column or beam with any general crosssection (input in
CUFSMstyleformat)usingninenodeS9R5thinshellfiniteelements.Theuserhasthe
abilitytoaddholesatspecificlocationsinthemember,dictatetheboundaryconditions
and application of load, specify the material properties, and impose imperfection,
residual stresses and plastic strains to define a members initial state. Input files for
eigenbucklinganalysisandnonlinearfiniteelementsimulationscanbegenerated.The
programwasusedthroughoutthisresearchtogenerategroupsofABAQUSinputfiles
forparameterstudies.Theprogramsetupusedtogeneratethenonlinearfiniteelement
modelsofthecolumnexperimentsisprovidedhere.

419

clear all
close all
sourceloc='C:\Documents and Settings\Cris\Desktop\cmoen\Cold Formed Steel - Holes
Research\Fall 2007\runbuck development\Rev_6NL\jhab'

%This example generates ABAQUS input files for the nonlinear


%simulation of 12 different columns with SSMA C-sections and evenly spaced
%slotted holes. The column boundary conditions are pinned-pinned warping free and the
column is
%loaded from both ends with a uniform compressive stress simulated as
%consistent nodal loads on the first two sets of cross-section nodes.
%25%CDF and 75% CDF imperfections are imposed on the member geometry with
%CUFSM local and distortional buckling mode shapes.
addpath([sourceloc
addpath([sourceloc
addpath([sourceloc
addpath([sourceloc
addpath([sourceloc
load
load
load
load

SSMAxsections
SSMAnames
SSMA_wvlengths
Ag

'\functions\filewriting\'])
'\functions\holes\'])
'\functions\'])
'\templates\'])
'\'])
%SSMA
%SSMA
%SSMA
%SSMA

cross section info


name list
cFSM wavelengths for Pcrl, Pcrd
gross cross-sectional area

%define the SSMA sections to create models for


sections=[12
86
11
73
39
95
72
56
47
75
66
87]
%define the imperfection magnitudes
imptypes=[25 75]
%define hole length (slotted holes considered here)
Lhole=4
%define rough hole spacing, will be adjusted in holes section of file
S=12
%define member lengths
Lc=[34
88
24
74

420

42
78
66
56
32
74
40
80]
%define hole depth such that Anet=0.70Ag
Anetfactor=0.7
count=1
for i=1:length(sections)
section_num=sections(i)
for j=1:length(imptypes)
%MEMBER LENGTH
L=Lc(i)
%MESH ALONG LENGTH
nele=L*2;
%NUMBER OF SECTION POINTS THROUGH THE THICKNESS
sectionpoints=5

RB2

%CROSS-SECTION DIMENSIONS
%
Z
%
%
A
%
X
% D2 /
I
\ D1
% RT2/_S2___
S
___S1_\RT1
%
\
|
/
% B2 \
|
/ B1
%
\
|
/
% ___F2_\__________________/_F1___
ABAQUS Y AXIS
%
RB2
H
RB1
%Dimensions are out-to-out, angles are in degrees, t is base metal +
%coating thickness, tbare is base metal thickness
%
[H
B1
B2
D1
D2
F1
F2
S1
S2
RT1
RT2
t
tbare]
dims=SSMAxsections(section_num,2:16)

RB1

%calculate hole depth


hhole=Ag(section_num)*(1-Anetfactor)./dims(15)
%CROSS-SECTION MESHING
%number of elements around the cross section
%[D1
RT1 B1 RB1 H RB2 B2 RT2 D2]
n=[2
2
2
2
16
2
2
2
2];
%CorZ=1 C-section, CorZ=2 Z-section
CorZ=1
[node,elem]=cztemplate(CorZ,dims,n)
nnodes=length(node(:,1)); %Number of FSM cross-section nodes
%Determine FE number of nodes and increment
nL=2*nele+1; %Number of FE nodes along the length
%Determine the node numbering increment along the length
if nnodes<100
FEsection_increment=100; %so along the length the numbering goes up by 100's
else
FEsection_increment=nnodes+1;
end
%ADD ADDITIONAL NODES
nodeadd=[]
%MATERIAL PROPERTIES

421

%steel
matprops(1).name='MAT100';
matprops(1).elastic=[29500 0.3];
matprops(1).plastic=[58.6, 0
64.1517, 0.00342827
68.2188, 0.00842827
72.0304, 0.0134283
77.9752, 0.0234283
82.2224, 0.0334283
85.7249, 0.0434283
88.4053, 0.0534283
90.7405, 0.0634283
92.652, 0.0734283
94.3657, 0.0834283
95.8299, 0.0934283
97.2001, 0.103428 ];
%IMPERFECTIONS
%*****IMPERFECTIONS*****
%type=0
no imperfections
%type=1
use mode shapes from ABAQUS results file
%type=2
input from file
%type 3
impose CUFSM shapes as imperfections
imperfections.type=3;
imperfections.filename=[];
imperfections.step=[];
imperfections.mode=[]
t=dims(15)
if imptypes(j)==25
if L>24
imperfections.magnitude=[0.14*t 0.64*t L/2000]
imperfections.wavelength=[SSMA_wvlengths(section_num,1)
SSMA_wvlengths(section_num,2) L]
else
imperfections.magnitude=[0.14*t 0.64*t]
imperfections.wavelength=[SSMA_wvlengths(section_num,1)
SSMA_wvlengths(section_num,2)]
end
elseif imptypes(j)==75
if L>24
imperfections.magnitude=[0.66*t 1.55*t L/1000]
imperfections.wavelength=[SSMA_wvlengths(section_num,1)
SSMA_wvlengths(section_num,2) L]
else
imperfections.magnitude=[0.66*t 1.55*t]
imperfections.wavelength=[SSMA_wvlengths(section_num,1)
SSMA_wvlengths(section_num,2)]
end
end
imperfections.plumb=[]
imperfections.member=[1] %1 for column, 2 for beam
%DEFINE HOLES
%Add holes to your member.
%hole.type=1 circular
%hole.type=2 rectangular
%hole.type=3 slotted w\radial ends
%hole.dimension=['width or length (ABAQUS x direction)' 'height or diameter']
%hole.location=['CUFSM cross section node (must be odd!)' 'longitudinal location'
'shift hole in direction of height']
%hole.thickness = thickness of finite elements making up hole, usually the same
as the rest of the member
%I've defined two slotted holes here in the web of the cross-section.
%number of holes
nhole=floor(L/S)
if nhole<1

422

nhole=1
end
%final hole spacing
Sfinal=floor(L/nhole)
spacing=Sfinal/2:Sfinal:L-Sfinal/2
hole.type=[3*ones(nhole,1)];
%define dimensions for slotted hole
hole.dimension=[Lhole*ones(nhole,1) hhole*ones(nhole,1)];
%define location of hole in cross-section
hole.location = [(length(node)+1)/2*ones(nhole,1) spacing' zeros(nhole,1)]
hole.thickness = [dims(1,15)*ones(length(hole.type),1)]
hole.material=[100*ones(length(hole.type),1)];
hole.groups=[100000+[1:length(hole.type)]];
hole.fill=[zeros(length(hole.type),1)];
%If you don't want holes, replace above with
%hole=[ ]
%MEMBER END LOADINGS
%Loading notation is similar to CUFSM. Apply P for compression, M for
%moment, or a combination of both. Compression at both ends of
%column are
%shown here. Loads are applied as consistent nodal loads in ABAQUS.
end1load.P=1;
end1load.Mxx=0;
end1load.Mzz=0;
end1load.M11=0;
end1load.M22=0;
end2load.P=-1;
end2load.Mxx=0;
end2load.Mzz=0;
end2load.M11=0;
end2load.M22=0;
%CALCULATE CONSISTENT NODAL LOADS ON MEMBER ENDS*****
unsymm=0
[end1cload, end2cload, A,
Ixx]=consist_endloads(node,elem,end1load,end2load,unsymm, nL, FEsection_increment);

%ABAQUS NODE SETS


%Define these node sets to apply boundary conditions in ABAQUS
%
nodesetinfo={'nodeset name' [xlim1 xlim2 xint] [ylim1 ylim2 yint] [zlim1
zlim2 zint] exclude}
%where nodes are grouped based on xlim1<=x<=xlim2 and ylim1<=y<=ylim2 and
zlim1<=z<=zlim2.
%Instead of ranges, assign xint,yint,zint to something other than zero to group
nodes at specific x,y,and z
%distance intervals
%
xlim1:xint:xlim2, ylim1:yint:ylim2, ylim1:yint:ylim2.
%The exclude command can be used to exclude previously defined node sets from the
current node set.
%exclude = 0
all nodes in range are included in nodeset
%exclude = m
excludes nodeset m from current nodeset
nodesetinfo={'ENDXZERO' [0 0 0] [-1000 1000 0] [-1000 1000 0] 0;
'ENDXL' [L L 0] [-1000 1000 0] [-1000 1000 0] 0;
'DISPDOF' [L L 0] [0 0 0] [0 0 0] 0;
'MID' [L/2 L/2 0] [max(node(:,3))-0.05 max(node(:,3))+0.05 0]
[max(node(:,2))/2-0.05 max(node(:,2))/2+0.05 0] 0};

%DEFINE SPRINGS
springs=[]
%DEFINE CONTACT SURFACES, NODE SURFACES, KINEMATIC CONSTRAINTS,....
surface.type={}
surface.type=[]
surface.local=[]

423

surface.coord=[]
surface.coupling={};
surface.interaction=[]
surface.contact=[]
surface.areadist=[]
%ABAQUS INP FILE NAME
jobname{count}=[SSMAnames{section_num} '_' num2str(i) '_' num2str(imptypes(j))];

%DEFINE ANALYSIS STEP


step(1).stepinfo={'STEP 1,' 'nlgeom, INC=' [260]};
step(1).solutiontype='Static, Riks';
step(1).solutionsteps={'0.25, ,1e-10, 1'};
step(1).solutioncontrols={ };
step(1).boundarycon={'ENDXZERO' 2 3;
'ENDXL' 2 3;
'MID' 1 1}
step(1).coupling=[]
step(1).loads={'*Cload' end1cload(:,1) 1 end1cload(:,2)./2;
'*Cload' end1cload(:,1)+200 1 end1cload(:,2)./2;
'*Cload' end2cload(:,1) 1 end2cload(:,2)./2;
'*Cload' end2cload(:,1)-200 1 end2cload(:,2)./2}
step(1).outrequest={'*Output, field, frequency=10';
'*Element Output';
'1,3,5';
'S,MISES';
'*Node Output';
'U';
'*Node Print, NSET=DISPDOF, SUMMARY=NO';
'U1,CF1'};
%WRITE ABAQUS INP FILE
%this is the important function, you can use this in for loops to generate
parameter studies
jhabnl(L, node, elem, nele, end1load, end2load, hole, nodesetinfo, surface,
nodeadd, step, jobname{count},matprops,imperfections,springs,sectionpoints)
count=count+1
end

end
%CREATE BEAST BATCH FILE
%Generates a linux batch file that will submit all of the parameter study
%.inp files to the queue manager on the beast.
ABQbeastscript(jobname,ones(length(jobname),1)*4,'cdmscript')
%Run the script at the beast command line with:
% bash cdmscript

424

Appendix B
ABAQUS element-based elastic buckling results
10B

ThisappendixcontainsthefiniteelementplatemodeldimensionsandABAQUScritical
elastic buckling stress results (fcr ) used in the Chapter 3 elastic buckling studies on
l

1832H

stiffenedandunstiffenedelements.

425

426

Stiffenedelementinuniaxialcompression,
transverselycenteredholes
Model
number

hole type

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75

circular
circular
circular
circular
circular
circular
circular
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted

hhole
in.
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50

h
in.
15.00
7.50
5.00
3.75
3.00
2.50
2.14
15.00
7.50
5.00
3.75
3.00
2.50
2.14
15.00
7.50
5.00
3.75
3.00
2.50
2.14
15.00
7.50
5.00
3.75
3.00
2.50
2.14
15.00
7.50
5.00
3.75
3.00
2.50
2.14
3.41
3.41
3.41
3.41
3.41
3.41
3.41
3.41
3.41
3.41
3.41
3.41
5.77
5.77
5.77
5.77
5.77
5.77
5.77
5.77
5.77
5.77
5.77
7.89
7.89
7.89
7.89
7.89
7.89
7.89
7.89
7.89
7.89
2.27
2.27
2.27
2.27
2.27
2.27
2.27

Lhole
in.
1.50
1.50
1.50
1.50
1.50
1.50
1.50
4.00
4.00
4.00
4.00
4.00
4.00
4.00
6.00
6.00
6.00
6.00
6.00
6.00
6.00
8.00
8.00
8.00
8.00
8.00
8.00
8.00
12.00
12.00
12.00
12.00
12.00
12.00
12.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00

427

S
in.
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
96.00
48.00
32.00
24.00
19.20
16.00
13.71
12.00
10.67
9.60
8.73
8.00
96.00
48.00
32.00
24.00
19.20
16.00
13.71
12.00
10.67
9.60
8.73
96.00
48.00
32.00
24.00
19.20
16.00
13.71
12.00
10.67
9.60
12.00
16.00
20.00
24.00
28.00
32.00
36.00

L
in.
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
12.00
16.00
20.00
24.00
28.00
32.00
36.00

t
in.
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346

hole
in.
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00

fcrl
ksi
0.57
2.21
4.91
8.88
14.21
20.61
28.03
0.53
1.99
4.38
8.01
13.39
20.68
28.09
0.49
1.80
3.96
7.27
12.44
20.73
28.13
0.44
1.65
3.67
6.86
12.06
20.78
28.21
0.37
1.49
3.36
6.53
11.83
20.80
22.82
9.95
9.95
9.95
9.94
9.94
9.92
9.91
9.86
9.94
9.77
9.72
9.80
3.29
3.29
3.29
3.28
3.25
3.28
3.18
3.17
3.26
3.23
3.02
1.84
1.84
1.83
1.81
1.80
1.77
1.81
1.81
1.68
1.71
26.47
25.43
25.02
24.86
24.80
24.77
24.75

76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145

slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
square
square
square
square
square
square
square

1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50

2.27
2.27
2.27
2.27
2.27
2.27
2.27
3.41
3.41
3.41
3.41
3.41
3.41
3.41
3.41
3.41
3.41
3.41
3.41
3.41
3.41
5.77
5.77
5.77
5.77
5.77
5.77
5.77
5.77
5.77
5.77
5.77
5.77
5.77
5.77
7.89
7.89
7.89
7.89
7.89
7.89
7.89
7.89
7.89
7.89
7.89
7.89
7.89
7.89
15.00
7.50
5.00
3.75
3.00
2.50
2.14
15.00
7.50
5.00
3.75
3.00
2.50
2.14
15.00
7.50
5.00
3.75
3.00
2.50
2.14

4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
1.50
1.50
1.50
1.50
1.50
1.50
1.50

40.00
44.00
48.00
60.00
72.00
84.00
96.00
12.00
16.00
20.00
24.00
28.00
32.00
36.00
40.00
44.00
48.00
60.00
72.00
84.00
96.00
12.00
16.00
20.00
24.00
28.00
32.00
36.00
40.00
44.00
48.00
60.00
72.00
84.00
96.00
12.00
16.00
20.00
24.00
28.00
32.00
36.00
40.00
44.00
48.00
60.00
72.00
84.00
96.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00

428

40.00
44.00
48.00
60.00
72.00
84.00
96.00
12.00
16.00
20.00
24.00
28.00
32.00
36.00
40.00
44.00
48.00
60.00
72.00
84.00
96.00
12.00
16.00
20.00
24.00
28.00
32.00
36.00
40.00
44.00
48.00
60.00
72.00
84.00
96.00
12.00
16.00
20.00
24.00
28.00
32.00
36.00
40.00
44.00
48.00
60.00
72.00
84.00
96.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00

0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0692
0.0692
0.0692
0.0692
0.0692
0.0692
0.0692
0.1038
0.1038
0.1038
0.1038
0.1038
0.1038
0.1038
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346

0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00

24.74
24.73
24.72
24.69
24.68
24.67
24.67
9.86
9.98
9.94
9.95
9.95
9.95
9.95
9.95
9.95
9.95
9.95
9.95
9.95
9.95
3.57
3.31
3.26
3.30
3.30
3.29
3.29
3.29
3.29
3.29
3.29
3.29
3.29
3.29
1.81
2.02
1.87
1.81
1.83
1.85
1.84
1.84
1.84
1.84
1.84
1.84
1.84
1.84
2.14
7.96
17.45
31.73
52.49
82.03
111.50
4.81
17.87
39.05
70.42
114.77
178.44
242.93
0.56
2.21
4.95
9.02
14.32
20.62
27.99

Stiffenedelementinuniaxialcompression,offsetholes

Model
number

hole type

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46

slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted

hhole

Lhole

hole

fcrl

in.
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50

in.
15.00
15.00
15.00
15.00
15.00
15.00
15.00
7.50
7.50
7.50
7.50
7.50
7.50
5.00
5.00
5.00
5.00
5.00
3.75
3.75
3.75
3.75
3.75
3.75
3.75
3.00
3.00
3.00
3.00
3.00
3.00
3.00
2.50
2.50
2.50
2.50
2.50
2.50
2.50
2.14
2.14
2.14
2.14
2.14
2.14
2.14

in.
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00

in.
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00

in.
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00

in.
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346

in.
0.00
0.94
1.88
2.81
3.75
4.69
5.63
0.00
0.47
0.94
1.41
1.88
2.34
0.00
0.31
0.63
0.94
1.25
0.07
0.14
0.20
0.27
0.34
0.41
0.47
0.05
0.09
0.14
0.18
0.23
0.27
0.32
0.03
0.06
0.09
0.12
0.15
0.18
0.21
0.02
0.04
0.06
0.08
0.10
0.12
0.14

ksi
0.53
0.54
0.54
0.54
0.54
0.54
0.54
1.99
1.98
1.96
1.96
1.91
1.85
4.38
4.17
3.79
3.71
3.38
7.81
7.43
7.03
6.65
6.31
5.99
5.70
12.65
11.76
10.95
10.23
9.59
9.01
8.49
20.12
18.55
17.15
15.91
14.80
13.81
12.91
28.54
28.54
27.99
25.88
23.98
22.26
20.72

429

Stiffenedelementinbending(Y=0.50h),transversely
centeredholes

Model
number

hole type

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28

slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted

hhole
in.
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50

h
in.
15.00
15.00
15.00
15.00
7.50
7.50
7.50
7.50
5.00
5.00
5.00
5.00
3.75
3.75
3.75
3.75
3.00
3.00
3.00
3.00
2.50
2.50
2.50
2.50
2.14
2.14
2.14
2.14

Lhole
in.
4.00
6.00
8.00
12.00
4.00
6.00
8.00
12.00
4.00
6.00
8.00
12.00
4.00
6.00
8.00
12.00
4.00
6.00
8.00
12.00
4.00
6.00
8.00
12.00
4.00
6.00
8.00
12.00

S
in.
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00

430

L
in.
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00

t
in.
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346

hole
in.
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00

Y
in.
7.50
7.50
7.50
7.50
3.75
3.75
3.75
3.75
2.50
2.50
2.50
2.50
1.88
1.88
1.88
1.88
1.50
1.50
1.50
1.50
1.25
1.25
1.25
1.25
1.07
1.07
1.07
1.07

fcrl
ksi
3.26
3.06
2.78
2.23
10.63
8.27
6.77
5.28
17.97
13.89
11.97
10.40
27.16
22.42
20.51
19.11
41.14
36.47
34.85
33.81
64.42
60.59
59.52
59.07
106.30
104.39
104.25
58.03

Stiffenedelementinbending(Y=0.50h),offsetholes
Model
number

hole type

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92

slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted

hhole
in.
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50

h
in.
15.00
15.00
15.00
15.00
15.00
15.00
15.00
7.50
7.50
7.50
7.50
7.50
7.50
5.00
5.00
5.00
5.00
5.00
3.75
3.75
3.75
3.75
3.75
3.75
3.75
3.00
3.00
3.00
3.00
3.00
3.00
3.00
2.50
2.50
2.50
2.50
2.50
2.50
2.50
2.14
2.14
2.14
2.14
2.14
2.14
2.14
15.00
15.00
15.00
15.00
15.00
15.00
15.00
7.50
7.50
7.50
7.50
7.50
7.50
5.00
5.00
5.00
5.00
5.00
3.75
3.75
3.75
3.75
3.75
3.75
3.75
3.00
3.00
3.00
3.00
3.00
3.00
3.00
2.50
2.50
2.50
2.50
2.50
2.50
2.50
2.14
2.14
2.14
2.14
2.14
2.14
2.14

Lhole
in.
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00

S
in.
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00

431

L
in.
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00

t
in.
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346

hole
in.
0.00
-0.94
-1.88
-2.81
-3.75
-4.69
-5.63
0.00
-0.47
-0.94
-1.41
-1.88
-2.34
0.00
-0.31
-0.63
-0.94
-1.25
-0.07
-0.14
-0.20
-0.27
-0.34
-0.41
-0.47
-0.05
-0.09
-0.14
-0.18
-0.23
-0.27
-0.32
-0.03
-0.06
-0.09
-0.12
-0.15
-0.18
-0.21
-0.02
-0.04
-0.06
-0.08
-0.10
-0.12
-0.14
0.00
0.94
1.88
2.81
3.75
4.69
5.63
0.00
0.47
0.94
1.41
1.88
2.34
0.00
0.31
0.63
0.94
1.25
0.07
0.14
0.20
0.27
0.34
0.41
0.47
0.05
0.09
0.14
0.18
0.23
0.27
0.32
0.03
0.06
0.09
0.12
0.15
0.18
0.21
0.02
0.04
0.06
0.08
0.10
0.12
0.14

Y
in.
7.50
7.50
7.50
7.50
7.50
7.50
7.50
3.75
3.75
3.75
3.75
3.75
3.75
2.50
2.50
2.50
2.50
2.50
1.88
1.88
1.88
1.88
1.88
1.88
1.88
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.25
1.25
1.25
1.25
1.25
1.25
1.25
1.07
1.07
1.07
1.07
1.07
1.07
1.07
7.50
7.50
7.50
7.50
7.50
7.50
7.50
3.75
3.75
3.75
3.75
3.75
3.75
2.50
2.50
2.50
2.50
2.50
1.88
1.88
1.88
1.88
1.88
1.88
1.88
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.25
1.25
1.25
1.25
1.25
1.25
1.25
1.07
1.07
1.07
1.07
1.07
1.07
1.07

fcrl
ksi
3.26
3.34
3.38
3.41
3.44
3.45
3.44
10.63
12.10
13.35
13.80
13.97
13.92
17.96
21.25
27.86
31.03
31.18
27.74
28.67
30.05
32.01
34.74
38.57
43.91
41.09
41.45
42.28
43.66
45.76
48.87
53.49
63.17
62.48
62.30
62.68
63.73
65.62
68.69
102.33
99.94
98.12
96.92
96.40
96.69
97.99
3.26
3.14
3.04
2.96
2.91
2.81
2.57
10.63
9.72
9.53
10.17
10.87
7.36
17.96
16.95
17.88
21.78
19.29
26.90
26.93
27.25
27.88
28.87
30.28
32.22
41.54
42.34
43.55
45.23
47.44
50.31
54.02
66.02
68.21
70.99
74.42
78.63
83.69
89.64
108.66
112.46
116.44
116.35
112.54
107.24
100.88

Stiffenedelementinbending(Y=0.75h),offsetholes
Model
number

hole type

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92

slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted

hhole
in.
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50

h
in.
15.00
15.00
15.00
15.00
15.00
15.00
15.00
7.50
7.50
7.50
7.50
7.50
7.50
5.00
5.00
5.00
5.00
5.00
3.75
3.75
3.75
3.75
3.75
3.75
3.75
3.00
3.00
3.00
3.00
3.00
3.00
3.00
2.50
2.50
2.50
2.50
2.50
2.50
2.50
2.14
2.14
2.14
2.14
2.14
2.14
2.14
15.00
15.00
15.00
15.00
15.00
15.00
15.00
7.50
7.50
7.50
7.50
7.50
7.50
5.00
5.00
5.00
5.00
5.00
3.75
3.75
3.75
3.75
3.75
3.75
3.75
3.00
3.00
3.00
3.00
3.00
3.00
3.00
2.50
2.50
2.50
2.50
2.50
2.50
2.50
2.14
2.14
2.14
2.14
2.14
2.14
2.14

Lhole
in.
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00

S
in.
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00

432

L
in.
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00

t
in.
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346

hole
in.
0.00
-0.94
-1.88
-2.81
-3.75
-4.69
-5.63
0.00
-0.47
-0.94
-1.41
-1.88
-2.34
0.00
-0.31
-0.63
-0.94
-1.25
-0.07
-0.14
-0.20
-0.27
-0.34
-0.41
-0.47
-0.05
-0.09
-0.14
-0.18
-0.23
-0.27
-0.32
-0.03
-0.06
-0.09
-0.12
-0.15
-0.18
-0.21
-0.02
-0.04
-0.06
-0.08
-0.10
-0.12
-0.14
0.00
0.94
1.88
2.81
3.75
4.69
5.63
0.00
0.47
0.94
1.41
1.88
2.34
0.00
0.31
0.63
0.94
1.25
0.07
0.14
0.20
0.27
0.34
0.41
0.47
0.05
0.09
0.14
0.18
0.23
0.27
0.32
0.03
0.06
0.09
0.12
0.15
0.18
0.21
0.02
0.04
0.06
0.08
0.10
0.12
0.14

Y
in.
11.25
11.25
11.25
11.25
11.25
11.25
11.25
5.63
5.63
5.63
5.63
5.63
5.63
3.75
3.75
3.75
3.75
3.75
2.81
2.81
2.81
2.81
2.81
2.81
2.81
2.25
2.25
2.25
2.25
2.25
2.25
2.25
1.88
1.88
1.88
1.88
1.88
1.88
1.88
1.61
1.61
1.61
1.61
1.61
1.61
1.61
11.25
11.25
11.25
11.25
11.25
11.25
11.25
5.63
5.63
5.63
5.63
5.63
5.63
3.75
3.75
3.75
3.75
3.75
2.81
2.81
2.81
2.81
2.81
2.81
2.81
2.25
2.25
2.25
2.25
2.25
2.25
2.25
1.88
1.88
1.88
1.88
1.88
1.88
1.88
1.61
1.61
1.61
1.61
1.61
1.61
1.61

fcrl
ksi
1.48
1.49
1.50
1.51
1.53
1.54
1.56
5.20
5.22
5.33
5.52
5.77
6.04
9.97
9.62
9.66
10.09
11.07
15.53
15.14
14.83
14.59
14.42
14.30
14.24
23.86
23.01
22.29
21.66
21.13
20.68
20.32
37.44
35.83
34.40
33.14
32.01
31.01
30.11
61.27
58.47
55.92
53.62
51.55
49.58
47.81
1.48
1.47
1.45
1.44
1.43
1.40
1.36
5.19
5.25
5.31
5.17
4.62
3.86
9.97
10.79
11.83
10.02
7.24
16.58
17.27
18.11
19.13
20.35
21.75
21.79
25.99
27.34
28.92
30.82
33.09
35.84
38.21
41.35
43.75
46.52
49.74
53.51
56.68
56.66
67.67
71.25
74.96
76.63
75.11
72.38
68.78

Unstiffenedelementinuniaxialcompression,
transverselycenteredholes
Model
number

hole type

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91

ang. slotted
ang. slotted
ang. slotted
ang. slotted
ang. slotted
ang. slotted
ang. slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
square
square
square
square
square
square
square
circular
circular
circular
circular
circular
circular
circular

hhole
in.
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50

h
in.
15.00
7.50
5.00
3.75
3.00
2.50
2.14
15.00
7.50
5.00
3.75
3.00
2.50
2.14
15.00
7.50
5.00
3.75
3.00
2.50
2.14
15.00
7.50
5.00
3.75
3.00
2.50
2.14
15.00
7.50
5.00
3.75
3.00
2.50
2.14
15.00
7.50
5.00
3.75
3.00
2.50
2.14
15.00
7.50
5.00
3.75
3.00
2.50
2.14
7.89
7.89
7.89
7.89
7.89
7.89
7.89
5.77
5.77
5.77
5.77
5.77
5.77
5.77
3.41
3.41
3.41
3.41
3.41
3.41
3.41
2.27
2.27
2.27
2.27
2.27
2.27
2.27
15.00
7.50
5.00
3.75
3.00
2.50
2.14
15.00
7.50
5.00
3.75
3.00
2.50
2.14

Lhole
in.
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
6.00
6.00
6.00
6.00
6.00
6.00
6.00
8.00
8.00
8.00
8.00
8.00
8.00
8.00
12.00
12.00
12.00
12.00
12.00
12.00
12.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50

433

S
in.
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
96.00
48.00
32.00
24.00
16.00
12.00
8.00
96.00
48.00
32.00
24.00
16.00
12.00
8.00
96.00
48.00
32.00
24.00
16.00
12.00
8.00
96.00
48.00
32.00
24.00
16.00
12.00
8.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00

L
in.
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
96.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00

t
in.
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0692
0.0692
0.0692
0.0692
0.0692
0.0692
0.0692
0.1038
0.1038
0.1038
0.1038
0.1038
0.1038
0.1038
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346

hole
in.
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00

fcrl
ksi
0.06
0.23
0.48
0.76
0.98
1.08
1.03
0.06
0.23
0.48
0.76
0.98
1.08
1.03
0.06
0.22
0.45
0.67
0.81
0.85
0.77
0.06
0.21
0.43
0.61
0.63
0.58
0.49
0.06
0.20
0.31
0.34
0.32
0.28
0.23
0.25
0.91
1.93
3.04
3.88
4.22
3.96
0.55
2.05
4.33
6.79
8.59
9.23
8.52
0.22
0.22
0.22
0.21
0.21
0.20
0.19
0.40
0.38
0.38
0.38
0.37
0.36
0.34
0.91
0.87
0.87
0.87
0.86
0.84
0.78
1.16
1.08
1.08
1.08
1.08
1.08
1.06
0.06
0.24
0.52
0.88
1.26
1.57
1.71
0.06
0.24
0.52
0.91
1.34
1.77
2.11

Unstiffenedelementinuniaxialcompression,offsetholes
Model
number

hole type

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92

slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted
slotted

hhole

Lhole

hole

fcrl

in.
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50

in.
15.00
15.00
15.00
15.00
15.00
15.00
15.00
7.50
7.50
7.50
7.50
7.50
7.50
5.00
5.00
5.00
5.00
5.00
3.75
3.75
3.75
3.75
3.75
3.75
3.75
3.00
3.00
3.00
3.00
3.00
3.00
3.00
2.50
2.50
2.50
2.50
2.50
2.50
2.50
2.14
2.14
2.14
2.14
2.14
2.14
2.14
15.00
15.00
15.00
15.00
15.00
15.00
15.00
7.50
7.50
7.50
7.50
7.50
7.50
5.00
5.00
5.00
5.00
5.00
3.75
3.75
3.75
3.75
3.75
3.75
3.75
3.00
3.00
3.00
3.00
3.00
3.00
3.00
2.50
2.50
2.50
2.50
2.50
2.50
2.50
2.14
2.14
2.14
2.14
2.14
2.14
2.14

in.
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00

in.
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00

in.
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00

in.
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346
0.0346

in.
0.00
-0.94
-1.88
-2.81
-3.75
-4.69
-5.63
0.00
-0.47
-0.94
-1.41
-1.88
-2.34
0.00
-0.31
-0.63
-0.94
-1.25
-0.07
-0.14
-0.20
-0.27
-0.34
-0.41
-0.47
-0.05
-0.09
-0.14
-0.18
-0.23
-0.27
-0.32
-0.03
-0.06
-0.09
-0.12
-0.15
-0.18
-0.21
-0.02
-0.04
-0.06
-0.08
-0.10
-0.12
-0.14
0.00
0.94
1.88
2.81
3.75
4.69
5.63
0.00
0.47
0.94
1.41
1.88
2.34
0.00
0.31
0.63
0.94
1.25
0.07
0.14
0.20
0.27
0.34
0.41
0.47
0.05
0.09
0.14
0.18
0.23
0.27
0.32
0.03
0.06
0.09
0.12
0.15
0.18
0.21
0.02
0.04
0.06
0.08
0.10
0.12
0.14

ksi
0.06
0.06
0.06
0.06
0.06
0.06
0.06
0.23
0.23
0.24
0.24
0.24
0.25
0.49
0.50
0.51
0.53
0.55
0.80
0.81
0.82
0.83
0.84
0.86
0.87
1.05
1.07
1.09
1.12
1.14
1.17
1.20
1.18
1.21
1.24
1.27
1.31
1.35
1.39
1.16
1.19
1.22
1.26
1.29
1.34
1.38
0.06
0.06
0.06
0.06
0.06
0.06
0.06
0.23
0.23
0.23
0.22
0.22
0.22
0.49
0.48
0.47
0.46
0.45
0.77
0.76
0.75
0.75
0.74
0.73
0.72
1.01
0.99
0.97
0.96
0.94
0.92
0.91
1.13
1.10
1.08
1.05
1.03
1.01
0.99
1.10
1.07
1.05
1.02
0.99
0.97
0.95

434

Appendix C
Derivation of elastic buckling coefficients for
unstiffened elements
1B

C.1kAforanunstiffenedelementincompression

The finite strip method is employed with CUFSM (Schafer and dny 2006) to

calculate the plate buckling coefficient for an unstiffened strip in compression, kA, as a
function of unstiffened strip aspect ratio (Lhole/hA) and the compressive stress ratio (A).
TheunstiffenedelementmodelsetupinCUFSMisprovidedinFigure C.1.Theresults
183H

from the CUFSM parameter study, where A is varied from 0 to 1, are presented in
FigureC.2.
1834H

Lhole
Simply
supported

f1

A =

hA

f2
f1

f2

free

Unstiffened Element

Section A-A

FigureC.1CUFSMfinitestripmodelingdefinitionforanunstiffenedelementincompression

ThefminsearchfunctioninMatlab(Mathworks2007)isusedtodeterminethevariables
1through5inthegeneralequationform:

435

kA =

0.578
+
A + 0.34

1 2 A

L
3 A + 4 + hole
hA

ThevariablesarechosentominimizethesumofthesquarederrorsbetweentheCUFSM
resultsinFigureC.2andthefittedcurve.Thecurvefittingresultsintheequation:
1835H

kA =

0.578
+
A + 0.34

2.70 1.76 A
L
0.024 A + 0.035 + hole
hA

10
9

plate buckling coeff., kA

8
7
6
5
4
3
2
1
0

5
Lhole/hA

10

FigureC.2TheplatebucklingcoefficientkAforanunstiffenedelementincompression(themultiplecurves
represent0A1withastepof0.1,11curvestotal)

The mean and standard deviation of the ABAQUS to predicted ratio when

0.1Lhole/yA2 is 1.14 and 0.61 respectively. As shown in Figure C.3, the accuracy of the
1836H

predictionisoftenconservativewithinthisaspectratiorange.For2<Lhole/yA10themean
andstandarddeviationoftheABAQUStopredictedratioare0.99and0.02respectively.

436

10
9

plate buckling coeff., kA

8
7
6
Fitted curve is
conservative predictor
of kA

5
4
3
2
1
0

0.2

0.4

0.6

0.8

1
1.2
Lhole/hA

1.4

1.6

1.8

FigureC.3ThefittedcurveforkAisaconservativepredictorwhenLhole/yA2

C.2 kB for an unstiffened element with compression


onthefreeedge

The finite strip method is employed with CUFSM (Schafer and dny 2006) to

calculatetheplatebucklingcoefficientforanunstiffenedelementwithcompressionon
thefreeedgeandtensiononthesimplysupportededge,kB,asafunctionofunstiffened
stripaspectratio(Lhole/hB)andthecompressivestressratio(B).Theunstiffenedelement
modelsetupinCUFSMisprovidedinFigureC.4.
1837H

437

Lhole
A

f1

B =

free

f2
f1

hA

f2

A
Simply
supported

Unstiffened Element

Section A-A

FigureC.4CUFSMfinitestripmodelingdefinitionforanunstiffenedelementwithcompressiononthefree
edge,tensiononthesimplysupportededge

TheresultsfromtheCUFSMparameterstudy,where Bisvariedfrom0to10,are

presentedinFigure C.5C.5.Astheportionoftheplatethatisintensionincreases(i.e.,
183H

1839H

Bincreases),thebucklingmodeswitchesfromonebuckledhalfwavetomultiplehalf
waves.

40
35

plate buckling coeff., kB

30
25

Buckles in several halfwavelength (=10 shown)

20
15
Buckles in one halfwavelength (=0 shown)

10
5
0

10

15

Lhole/hB

FigureC.5VariationinplatebucklingcoefficientkBforanunstiffenedelementwithBrangingfrom0to10

ApolynomialcurveisfittotheminimumkBasshowninFigureC.6:
1840H

k B = 0.340 B2 + 0.100 B + 0.573 .

438

ThemeanandstandarddeviationoftheCUFSMminimatofittedcurvepredictionratio
are1.03and0.11respectively.

40
CUFSM
Fitted curve

35

plate buckling coeff., kB

30
25
20
15
10
5
0

10

FigureC.6CurvefittominimumkBforBrangingfrom0to10

ThisapproximationisaccuratewhenLhole/yB>5butdoesnotcapturetheboostinkBwhen
Lhole/hBissmall.SinceLhole/hBmayoftenbelessthan1whenconsideringcommonplateand
hole sizes, it is important to have a viable estimate of kB to avoid overly conservative
predictions. A family of curves is fit to the CUFSM predictions for the case when
0.25Lhole/yB2andwhereBisvariedfrom0to10,resultinginthefollowingequation:
2

h
0.38 B + 1.6 B + 0.49
L
Lhole
, 0 hole 2, 0 B 10
kB =
0.1
hB
h
0.3
0.20 B + B + 0.14
Lhole
1.8

439

The equation provides an accurate representation of kB as demonstrated in Figure C.7.


184H

The mean and standard deviation of the CUFSM to predicted ratio are 0.01 and 0.03
respectively.

60

50

kB

40

30

20

10

0.5

1.5
Lhole/hB

2.5

FigureC.7FamilyofcurvesusedtosimulateboostinkBwhenLhole/hB2,Brangesfrom0to10

440

Appendix D
Elastic buckling prediction method of crosssectional elements with holes
12B

B1 Critical Elastic Buckling Stress of Elements with Perforations


B1.1 Uniformly Compressed Stiffened Element
For uniformly compressed stiffened elements with uniformly spaced perforations
satisfying the limits

S 1.5 and S 2 ,
h

L hole

the critical elastic buckling stress, fcrl, is

f crl = min[f cr , f crh ] .

(Eq. B1.1-1)

The critical elastic buckling stress, fcr, without the influence of perforations is
2

f cr = k

2E t ,

12(1 2 ) h

(Eq. B1.1-2)

where k=4 for a stiffened element with L/h>4.


The critical elastic buckling stress, fcrh, with the influence of perforations is

f crh = f crh ,net (1 h hole h ) ,

(Eq. B1.1-3)

where the critical elastic buckling stress, fcrh,net, at the location of a perforation is

f crh , net = min[f crA , f crB ] .

(Eq. B1.1-4)

The critical elastic buckling stress, fcri,of unstiffened strip i is


2

f cri = k i

2 E t and i = A or B,

12(1 2 ) h i

(Eq. B1.1-5)

where

L hole h i 1 , k i = 0.425 +

(L hole

0.2

hi )

0.95

(Eq. B1.1-6)

0.6

441

L hole h i < 1 , k i = 0.925 , and i = A or B.

(Eq. B1.1-7)

B1.2 Stiffened Element Under Stress Gradient

For stiffened elements under a stress gradient with uniformly spaced perforations
satisfying the limits

S 1.5 and S 2 ,
h

L hole

the critical elastic buckling stress, fcrl, is

f crl = min[f cr , f crh ] .

(Eq. B1.2-1)

The critical elastic buckling stress, fcr, without the influence of perforations is
2

f cr = k

2E t ,

12(1 2 ) h

(Eq. B1.2-2)

where

k = 4 + 2(1 + ) + 2(1 + )
3

(Eq. B1.2-3)

and

= f 2 f1 = (h Y ) Y .

(Eq. B1.2-4)

The critical elastic buckling stress, fcrh, with the influence of perforations is

h A , and
for hA+hhole Y, f = f
crh
crh ,net (1 + A )
Y

(Eq. B1.2-5)

h hole
h
for hA+hhole < Y, f = f
2 A hole ,
crh
crh , net 1

(Eq. B1.2-6)

where

A =

Y hA .
Y

Eq. (B1.2-7)

The critical elastic buckling stress, fcrh,net, at the location of a perforation is

f crh ,net = min[f crA , f crB ]

(Eq. B1.2-8)

Consideration of unstiffened strip A buckling is required only if hA<Y,

f crA

2E t

= kA
12(1 2 ) h A

(Eq. B1.2-9)

where

442

kA =

0.578

A + 0.34

2.70 1.76 A
2
0.024 A + 0.035 + (L hole h A )

(Eq. B1.2-10)

Consideration of unstiffened strip B buckling is required only if hA+hhole<Y,

f crB = k B

2E t

12(1 2 ) h B

Y h A h hole

(Eq. B1.2-12)

where
for Lhole/hB>2

k B = 0.340 B2 + 0.100 B + 0.573 ,

(Eq. B1.2-12)

for Lhole/hB2
2

h
+ 1.6 B + 0.49
L hole
kB =
0.1
h
0.3
0.20 B + B + 0.14
L hole
0.38 B

1.8

(Eq. B1.2-13)

and

B =

h Y
, 0 B 10 .
Y hA h hole

(Eq. B1.2-14)

443

B1.3 Unstiffened Element Under Uniform Compression with Perforations

For uniformly compressed unstiffened elements with uniformly spaced perforations


satisfying the limits

h hole
S
L hole
L
0.50 , and
2
10 , hole 10 ,
h
L hole
hA
hB

(Eq. B1.3-1)

the critical elastic buckling stress, fcrl, is

f crl = min[f cr , f crh ] .

(Eq. B1.3-2)

The critical elastic buckling stress, fcr, without the influence of perforations is
2

f cr = k

2E t ,

12 (1 2 ) h

(Eq. B1.3-3)

where k=0.425 for unstiffened elements with L/h>4.


The critical elastic buckling stress, fcrh, with the influence of perforations is
2

2E t
h
, f crA 1 hole ,
f crh = min k

2
h

12 1 h

(Eq. B1.3-4)

where

L
k = 0.4251 0.062 hole
hA

(Eq. B1.3-5)

and fcrA is calculated with Eq. B1.1-5.

444

Appendix E
Derivation of global critical elastic buckling load for
a column with holes
13B

Thisderivationdevelopstheequationfortheflexuralcriticalelasticbucklingloadof

a column with two holes spaced symmetrically about the longitudinal midline of a
column.Theconclusionsofthisderivationareusedasthefoundationfortheweighted
properties approach for approximating Pcre for columns and beams with holes as
describedinSection 4.2.7.3.1.1.INHisthemomentofinertiaofthecolumncrosssection
1842H

awayfromtheholeandIHisthemomentofinertiaatthehole.

INH
IH

INH

l1

v
l2
l3

l4
L

IH
INH
FigureE.1Longcolumnwithtwoholesspacedsymmetricallyaboutthelongitudinalmidline.

A conservation of energy approach is employed in this derivation, and specifically

the RayleighRitz Method. The derivation is founded on the fundamental principle


relatingthestrainenergyandpotentialenergyofthecolumn,UandWrespectively:

445

= (U W ) = 0
where the column strain energy U is:
2

d 2v
1
U = EI 2 dx
2
dx
and the column potential energy is
2

1 dv
W = P dx .
2 dx
Applying the RaleighRitz method, we assume a shape function in the deformed
(buckled)configurationofthecolumn:

v (x ) = sin

x
L

The derivatives of this function are calculated:


dv

x
= cos
dx
L
L
2
2
d v

x
= 2 sin
2
dx
L
L
andthensubstitutedintotheequationsforUandWwhicharewrittenalongthelength
ofthecolumnas:

EI
EI H l 2 2 4
x
2 x
+
sin
dx
sin 2 dx + NH

4
0 L4

l
2 1 L
2
L
L
EI l 4 2 4
EI H L 2 4
x
2 x
sin
dx
sin 2 dx
+ H
+

4
4

2 l3 L
2 l4 L
L
L

U=

EI NH
2

l1

2 4

and
W=

1 L 2 2
P 2 2
2 x
=
P
cos
dx
.

2 0 L2
4L
L

446

l3

l2

2 4
4

x
sin 2 dx
L

ThederivativeofUistakenwithrespecttothearbitraryshapefunctionamplitude:
l3
L
dU EI NH 4 l1 2 x
x
x
sin dx + sin 2 dx + sin 2 dx
=
4

l2
l4
d
L
L
L
L
0
l4
EI H 4 l 2 2 x
x
+
+ sin dx + sin 2 dx

4
l
l
3
L
L
L
1

The definite integrals inside the derivative are then solved resulting in:
dU EI NH 4
=
d
L4
+

l 1 l 3 l 2 L l 4 L
2l 3
2l 1
2l 2
2l 4
2 + 2 2 + 2 2 + 4 sin L sin L + sin L + sin L

2l 3
EI H 4 LH
2l 2
2l 1
2l 4
L
+
+ sin
sin
+ sin
sin

4
L
L
L
L
L
2 4

Thelengthtermsfromtheintegrationsumtothelengthofcolumnwithoutahole,LNH,
and the length of column with a hole, LH. When the holes are symmetric about the
longitudinalmidlineofthecolumn,thetrigonometrictermscancelasshowninFigure
E.2andtheequationabovesimplifiesto:

dU EI NH 4 LNH EI NH 4 LH
=
+
2
2
d
L4
L4

1.5

H trig term
NH trig term

sin(2x/L)

0.5

-0.5

-1

-1.5

Trig terms cancel for holes symmetric


about longitudinal midline
0

0.1

0.2

0.3

0.4

0.5
x/L

0.6

0.7

0.8

0.9

FigureE.2Trigonometrictermsinenergysolutioncancelwhenholesaresymmetricaboutlongitudinal
midlineofcolumn.

447

Thesolutionforthepotentialenergyofthecolumnisnotdependentuponthemoment
ofinertiaandthereforethederivativecanbesolveddirectlyas:

dW P 2
=
d
2L
TheloadPthatminimizesthevariationinenergyisthecriticalelasticbucklingload,Pcre:

= (U W ) =

dU dW

= 0
d d

Equatingthevariationalenergyterms:

EI NH 4 LNH EI H 4 LH P 2
+

=0
L4
L4
2
2
2L
resultsinasolutionforPcrewherethemomentofinertiaisaweightedaverageofthenet
andgrosscrosssectionofthecolumn.

Pcre =

2 E I NH LNH + I H LH

L2

Pcreforacolumnwiththegeneralcaseofi=1..nholescanbeapproximatedas:

Pcre =

2 E I NH (LNH + TNH ) + I H (LH + TH )

L2

where
TNH =

L
2

2Lc ,i Lhole ,i
L
sin
, TH =
cos

2
i =1
L L
n

2Lc ,i Lhole ,i
sin
L L

cos
i =1

Lc,iisthedistancefromtheendofthecolumntothecenterlineofholeiandLhole,iisthe
lengthofholei.

448

Appendix F
Column experiment results
14B

449

ColumnSpecimen362124NH

362-1-24-NH

362-1-24-NH
Flange displacement (inches)

Column axial load (kips)

-14
-12
-10
-8
-6
Ptest=10.48 kips

-4
-2
0

0.05
0.1
0.15
0.2
Column axial displacement (inches)

1.5
W

0.5
0
-0.5
-1
-1.5

0.05
0.1
0.15
0.2
Column axial displacement (inches)

Notes:

Loaded N to S instead of S to N. Adjusted all geometry measurements.


Lips rotated and not touching bottom platen after peak load.

450

ColumnSpecimen362224NH

362-2-24-NH

362-2-24-NH
Flange displacement (inches)

Column axial load (kips)

-14
-12
-10
-8
-6
Ptest=10.51 kips

-4
-2
0

0.05
0.1
0.15
0.2
Column axial displacement (inches)

Notes:

451

1.5
1
0.5
0
W

-0.5

-1
-1.5

0.05
0.1
0.15
0.2
Column axial displacement (inches)

ColumnSpecimen362324NH

362-3-24-NH

362-3-24-NH
Flange displacement (inches)

Column axial load (kips)

-14
-12
-10
-8
-6
Ptest=10.15 kips

-4
-2
0

0.05
0.1
0.15
0.2
Column axial displacement (inches)

1.5
W

0.5
0
-0.5
-1
-1.5

0.05
0.1
0.15
0.2
Column axial displacement (inches)

Notes:

Bottom lips rotated at 7 kips post-peak and are not bearing on platen.

452

ColumnSpecimen362124H

362-1-24-H

362-1-24-H
Flange displacement (inches)

Column axial load (kips)

-14
-12
-10
-8
-6
Ptest=10.01 kips

-4
-2
0

0.05
0.1
0.15
0.2
Column axial displacement (inches)

1.5
1
0.5

-0.5
-1
-1.5

Visible buckling of web on either side of hole at 7 kips.

453

Notes:

0.05
0.1
0.15
0.2
Column axial displacement (inches)

ColumnSpecimen362224H
Local buckling at hole
(unstiffened strip)

(a) P=0 kips

(b) P=10.4 kips


(peak load)

(c) P=7.0 kips

362-2-24-H

362-2-24-H
Flange displacement (inches)

Column axial load (kips)

-14
-12
-10
-8
-6
Ptest=10.38 kips

-4
-2
0

0.05
0.1
0.15
0.2
Column axial displacement (inches)

1.5
1
0.5

0
-0.5
-1
-1.5

Notes:

Visible buckling of web on either side of hole at 5 kips.

454

0.05
0.1
0.15
0.2
Column axial displacement (inches)

ColumnSpecimen362324H

362-3-24-H

362-3-24-H
Flange displacement (inches)

Column axial load (kips)

-14
-12
-10
-8
-6
Ptest=9.94 kips

-4
-2
0

0.05
0.1
0.15
0.2
Column axial displacement (inches)

1.5
1
0.5
0
W

-0.5

-1
-1.5

0.05
0.1
0.15
0.2
Column axial displacement (inches)

Notes:

Accidental preload to 3 kips when adjusting specimen for test.

455

ColumnSpecimen362148NH

362-1-48-NH

362-1-48-NH
Flange displacement (inches)

Column axial load (kips)

-14
-12
-10
-8
-6
Ptest=9.09 kips

-4
-2
0

0.05
0.1
0.15
0.2
Column axial displacement (inches)

1.5
W

0.5
0
-0.5
-1
-1.5

0.05
0.1
0.15
0.2
Column axial displacement (inches)

Notes:

Good end conditions no visible gaps.


9 kips a metallic noise yielding of west flange and increase in local wavelengths.
Column failed by global-torsional collapse.

456

ColumnSpecimen362248NH
Peak Load

362-2-48-NH

362-2-48-NH
Flange displacement (inches)

Column axial load (kips)

-14
-12
-10
-8
-6
Ptest=9.49 kips

-4
-2
0

0.05
0.1
0.15
0.2
Column axial displacement (inches)

Notes:

Tight end conditions at 1.5 kips.


Local buckling at 6.5 kips .
No sounds for this test.
Column failed by global-torsional collapse.

457

1.5
W

0.5
0
-0.5
-1
-1.5

0.05
0.1
0.15
0.2
Column axial displacement (inches)

ColumnSpecimen362348NH
Peak Load

362-3-48-NH

362-3-48-NH
Flange displacement (inches)

Column axial load (kips)

-14
-12
-10
-8
-6
Ptest=9.48 kips

-4
-2
0

0.05
0.1
0.15
0.2
Column axial displacement (inches)

Notes:

Local buckling first observed at 6.5 kips.


Local wavelengths lengthen at 8.5 kips.
Yielding of flange lips at 9 kips (near peak).
Column failed by global-torsional collapse.

458

1.5
W

0.5
0
-0.5
-1
-1.5

0.05
0.1
0.15
0.2
Column axial displacement (inches)

ColumnSpecimen362148H
Peak Load

362-1-48-H

362-1-48-H
Flange displacement (inches)

Column axial load (kips)

-14
-12
-10
-8
-6
Ptest=8.95 kips

-4
-2
0

0.05
0.1
0.15
0.2
Column axial displacement (inches)

Notes:

No visible gaps and ends under 1 kip.


Local buckling is visible at 7 kips.
Local half-waves merge at 8.5 kips.
Bulging of web at hole occurs near peak load.
Column failed by global-torsional collapse.

459

1.5
W

0.5
0
-0.5
-1
-1.5

0.05
0.1
0.15
0.2
Column axial displacement (inches)

ColumnSpecimen362248H
Peak Load

362-2-48-H

362-2-48-H
Flange displacement (inches)

Column axial load (kips)

-14
-12
-10
-8
-6
Ptest=9.18 kips

-4
-2
0

0.05
0.1
0.15
0.2
Column axial displacement (inches)

1.5
1
0.5
0
-0.5
-1
-1.5

W
E

0.05
0.1
0.15
0.2
Column axial displacement (inches)

Notes:

End conditions tight at 4 kips.


Local buckling visible at 6.5 kips.
Distortional buckling seems to increase as load-displacement softens.
East LVDT reaches limit of range as column starts to twist.
Column failed by global-torsional collapse.

460

ColumnSpecimen362348H
Peak Load

362-3-48-H

362-3-48-H
Flange displacement (inches)

Column axial load (kips)

-14
-12
-10
-8
-6
Ptest=9.37 kips

-4
-2
0

0.05
0.1
0.15
0.2
Column axial displacement (inches)

Notes:

Local buckling visible at 6.5 kips.


Column failed by global-torsional collapse.

461

1.5
W

0.5
0
-0.5
-1
-1.5

0.05
0.1
0.15
0.2
Column axial displacement (inches)

ColumnSpecimen600124NH
Peak Load

600-1-24-NH

600-1-24-NH
Flange displacement (inches)

Column axial load (kips)

-14
-12
-10
-8
-6
Ptest=11.93 kips

-4
-2
0

0.05
0.1
0.15
0.2
Column axial displacement (inches)

Notes:

Local and distortional waves seem to stay separate.


8 kips (post-peak) east flange buckles.

462

1.5
1
0.5
0
W

-0.5

-1
-1.5

0.05
0.1
0.15
0.2
Column axial displacement (inches)

ColumnSpecimen600224NH
Peak Load

600-2-24-NH

600-2-24-NH
Flange displacement (inches)

Column axial load (kips)

-14
-12
-10
-8
-6
Ptest=11.95 kips

-4
-2
0

0.05
0.1
0.15
0.2
Column axial displacement (inches)

1.5
W

0.5
0
-0.5
-1
-1.5

0.05
0.1
0.15
0.2
Column axial displacement (inches)

Notes:

Web has large curve when placing specimen on bottom platen.


Visible gap between platen and specimen at top west web-flange corner - 5 kips.
10 kips (post-peak)- flanges buckle and lose contact with bottom platen.

463

ColumnSpecimen600324NH
Peak Load

600-3-24-NH

600-3-24-NH
Flange displacement (inches)

Column axial load (kips)

-14
-12
-10
-8
-6
Ptest=12.24 kips

-4
-2
0

0.05
0.1
0.15
0.2
Column axial displacement (inches)

1.5
W

0.5
0
-0.5
-1
-1.5

0.05
0.1
0.15
0.2
Column axial displacement (inches)

Notes:

Specimen failed at bottom end condition, web rolled over and was not bearing on platen.

464

ColumnSpecimen600124H
Peak Load

600-1-24-H

600-1-24-H
Flange displacement (inches)

Column axial load (kips)

-14
-12
-10
-8
-6
Ptest=12.14 kips

-4
-2
0

0.05
0.1
0.15
0.2
Column axial displacement (inches)

1.5
1
0.5
0
W

-0.5

-1
-1.5

0.05
0.1
0.15
0.2
Column axial displacement (inches)

Notes:

Slight gap at east top web-flange corner - 3 kips, gap is closed at 11 kips.
East flange gives way at 11 kips with dip in load-disp. curve, may be related to above.
Loud popping sound at 8 kips (post-peak) and large change in load-displ. slope.

465

ColumnSpecimen600224H
Peak Load

600-2-24-H

600-2-24-H
Flange displacement (inches)

Column axial load (kips)

-14
-12
-10
-8
-6
Ptest=11.62 kips

-4
-2
0

0.05
0.1
0.15
0.2
Column axial displacement (inches)

1.5
W

0.5
0
-0.5
-1
-1.5

0.05
0.1
0.15
0.2
Column axial displacement (inches)

Notes:

Specimen failure mode similar to that of a no-hole specimen.

466

ColumnSpecimen600324H
Peak Load

600-3-24-H

600-3-24-H
Flange displacement (inches)

Column axial load (kips)

-14
-12
-10
-8
-6
Ptest=11.79 kips

-4
-2
0

0.05
0.1
0.15
0.2
Column axial displacement (inches)

Notes:

Good contact with platens.

467

1.5
W

0.5
0
-0.5
-1
-1.5

0.05
0.1
0.15
0.2
Column axial displacement (inches)

ColumnSpecimen600148NH
Peak Load

600-1-48-NH

600-1-48-NH
Flange displacement (inches)

Column axial load (kips)

-14
-12
-10
-8
-6
Ptest=11.15 kips

-4
-2
0

0.05
0.1
0.15
0.2
Column axial displacement (inches)

1.5
W

0.5
0
-0.5
-1
-1.5

0.05
0.1
0.15
0.2
Column axial displacement (inches)

Notes:

Local buckling first observed at 4.5 kips (11 half-waves).


Distortional wave becomes prominent at 10 kips.
Loud noises 1 minute apart L waves turn to D waves at north, then south ends.

468

ColumnSpecimen600248NH
Peak Load

600-2-48-NH

600-2-48-NH
Flange displacement (inches)

Column axial load (kips)

-14
-12
-10
-8
-6
Ptest=11.44 kips

-4
-2
0

0.05
0.1
0.15
0.2
Column axial displacement (inches)

1.5
W

0.5
0
-0.5
-1
-1.5

0.05
0.1
0.15
0.2
Column axial displacement (inches)

Notes:

Gap between platen and specimen at top east flange-web corner closes at 2 kips.
Can see distortional shape developing at 4.5 kips.
Local buckling visible at 5 kips.
Two loud bangs (peak load, 10.5 kips post peak) local web waves change to D waves.
Flange distortion slows at 7 kips post-peak.
469

ColumnSpecimen600348NH
Peak Load

600-3-48-NH

600-3-48-NH
Flange displacement (inches)

Column axial load (kips)

-14
-12
-10
-8
-6
Ptest=11.29 kips

-4
-2
0

0.05
0.1
0.15
0.2
Column axial displacement (inches)

1.5
W

0.5
0
-0.5
-1
-1.5

0.05
0.1
0.15
0.2
Column axial displacement (inches)

Notes:

Gap between platen and specimen at east top flange closes at 1 kip.
Loud sound at peak load L waves change to D waves in web.

470

ColumnSpecimen600148H
Peak Load

600-1-48-H

600-1-48-H
Flange displacement (inches)

Column axial load (kips)

-14
-12
-10
-8
-6
Ptest=11.16 kips

-4
-2
0

0.05
0.1
0.15
0.2
Column axial displacement (inches)

1.5
W

0.5
0
-0.5
-1
-1.5

0.05
0.1
0.15
0.2
Column axial displacement (inches)

Notes:

Local buckling and DH mode visible at 5 kips.


Loud noise at 9.5 kips L waves changes to D wave in web.

471

ColumnSpecimen600248H
Peak Load

600-2-48-H

600-2-48-H
Flange displacement (inches)

Column axial load (kips)

-14
-12
-10
-8
-6
Ptest=11.7 kips

-4
-2
0

0.05
0.1
0.15
0.2
Column axial displacement (inches)

1.5
W

0.5
0
-0.5
-1
-1.5

0.05
0.1
0.15
0.2
Column axial displacement (inches)

Notes:

Local buckling visible at 4 kips.


D wave interrupted by large crease.
L web waves change to D waves (9 kips post-peak, 7.5 kips post-peak)

472

ColumnSpecimen600348H
Peak Load

600-3-48-H

600-3-48-H
Flange displacement (inches)

Column axial load (kips)

-14
-12
-10
-8
-6
Ptest=11.16 kips

-4
-2
0

0.05
0.1
0.15
0.2
Column axial displacement (inches)

Notes:

Good platen bearing conditions.


Loud noise at 7.5 kips post-peak.
Yielding in the west flange first, then east flange.

473

1.5
W

0.5
0
-0.5
-1
-1.5

0.05
0.1
0.15
0.2
Column axial displacement (inches)

Appendix G
Residual stresses backstress for kinematic
hardening implementation
15B

Implementation of a kinematic hardening rule requires that the center of the yield

surface,instressspace,beknownforanymaterialwhichhasbeenyieldedpriortothe
loading of interest. The coordinates of the center of the yield surface (1, 2, 3),
knownasthebackstress,cannotbedirectlycalculatedfromthestressesderivedherein
because work hardening was ignored in the residual stress derivations. However, the
plastic strains developed in the manufacturing process provide a means by which the
backstressmaybeapproximated,asprovidedinthisappendix.

Thegeneralequationforeffectivestressisdefinedas

e =

1
2

( 1 2 )2 + ( 2 3 )2 + ( 3 1 )2 .

(G.1)

Giventhatthethroughthicknesssheetstressesarezero(2=0),Eq.(G.1)reducesto

e = 1 2 1 3 + 3 2

(G.2)

Consider the contribution to the backstress that develops due to coiling. From Eq.

(6.18) we know the plastic strain, pcoiling. With pcoiling and knowing the material stress
1843H

strain relation (i.e., Figure 6.23) the effective stress at that plastic strain, eycoiling maybe
184H

determined.ConsistentwiththeresidualstressderivationofEq. (6.8),weassume =0.3


1845H

and

474

1coiling = 3coiling .

(G.3)

Finally,substitutingtheprecedingintoEq.(G.2)resultsin

coiling
3

eycoiling

Similarly for cold bending the corners, from Eq. (6.21) we know the plastic strain,

2 + 1

(G.4)

1846H

pbend. With pbend and knowing the material stressstrain relation (i.e., Figure 6.23), we
1847H

determine the effective stress at that plastic strain, eybend. Consistent with the residual
stressderivationofEq.(6.12),weassume=0.5and
184H

3bend = 1bend ,

(G.5)

andthusfind

bend
1

eybend
2 + 1

(G.6)

Thebackstressisthendeterminedas:

1 = 1coiling + 1bend yield

2 = 0

3 = 3coiling + 3bend yield ,

(G.7)

where yield isthevirginyieldstressofthesteel.Thisestimateassumesthatthechanges


in material properties from coiling, uncoiling, and flattening and coldforming do not
influenceoneanother.

475

Appendix H
Experiment true stress-strain curves
16B

The average true plastic stressstrain curves are provided here for each of the 24

columntestsreportedin Chapter5.Foreachspecimen,threeengineeringstressstrain
1849H

curves(westflange,eastflange,andweb)wereaveragedandthentransformedintotrue
stressesandstrainswiththefollowingequations:

true = ln(1 + o )

true = o (1 + o )
true and true are the true stress and strain and o and o are the engineering stress and
strain in the above equations. The tables in this appendix provide just the plastic
componentofthetruestrainsincethisiswhatisrequiredinABAQUS:

p = true yield ,where yield =

yield
E

The true stressstrain curves presented here were modified prior to their
implementationinABAQUStoensureplasticityinitiatedattheyieldstressandnot
theproportionallimit.RefertoSection7.2.1.4fordetailsonthismodelingdecision.
1850H

476


Specimen362124NH,362224NH,362324NH

100
90
80

true stress, ksi

70
60
50
40
30

Avg. Yield Stress=55.1 ksi

20
10
0

0.02

0.04
0.06
0.08
true plastic strain
true plastic
strain, p

true stress,
true

0
0.001
0.002
0.007
0.012
0.017
0.027
0.037
0.047
0.057
0.067

ksi
33.05
46.10
51.88
60.30
64.89
68.37
73.97
78.12
81.27
83.83
86.16

477

0.1

0.12

Specimen362124H

100
90
80

true stress, ksi

70
60
50
40
30

Avg. Yield Stress=57.9 ksi

20
10
0

0.02

0.04
0.06
0.08
true plastic strain
true plastic
strain, p

true stress,
true

0
0.001
0.002
0.007
0.012
0.017
0.027
0.037
0.047
0.057
0.067
0.077
0.087
0.097
0.107

ksi
34.2
50.0
56.1
64.4
68.3
72.0
78.6
82.5
86.2
88.7
91.0
92.9
94.6
96.2
97.5

478

0.1

0.12

Specimen362224H

100
90
80

true stress, ksi

70
60
50
40
30

Avg. Yield Stress=57.1 ksi

20
10
0

0.02

0.04
0.06
0.08
true plastic strain
true plastic
strain, p

true stress,
true

0
0.001
0.002
0.007
0.012
0.017
0.027
0.037
0.047
0.057
0.067
0.077
0.087
0.097
0.107

ksi
27.7
45.1
53.3
62.8
67.1
71.2
76.7
81.3
84.7
87.4
89.7
91.6
93.4
94.8
96.2

479

0.1

0.12

Specimen362324H

100
90
80

true stress, ksi

70
60
50
40
30

Avg. Yield Stress=56 ksi

20
10
0

0.02

0.04
0.06
0.08
true plastic strain
true plastic
strain, p

true stress,
true

0
0.001
0.002
0.007
0.012
0.017
0.027
0.037
0.047
0.057
0.067
0.077
0.087
0.097
0.107

ksi
31.7
47.8
53.6
61.1
64.7
68.4
74.8
78.6
82.2
84.6
86.9
88.8
90.4
91.9
93.1

480

0.1

0.12

Specimen362148NH

100
90
80

true stress, ksi

70
60
50
40
30

Avg. Yield Stress=59.7 ksi

20
10
0

0.02

0.04
0.06
0.08
true plastic strain
true plastic
strain, p

true stress,
true

0
0.001
0.002
0.007
0.012
0.017
0.027
0.037
0.047
0.057
0.067
0.077
0.087
0.097
0.107

ksi
37.8
51.6
56.6
64.3
68.2
72.0
78.1
82.1
85.9
88.3
90.8
92.5
94.3
95.7
97.1

481

0.1

0.12

Specimen362248NH

100
90
80

true stress, ksi

70
60
50
40
30

Avg. Yield Stress=59.2 ksi

20
10
0

0.02

0.04
0.06
0.08
true plastic strain
true plastic
strain, p

true stress,
true

0
0.001
0.002
0.007
0.012
0.017
0.027
0.037
0.047
0.057
0.067
0.077
0.087
0.097
0.107

ksi
35.0
50.9
56.5
64.3
68.3
72.1
78.2
82.3
86.1
88.5
90.9
92.8
94.5
96.0
97.3

482

0.1

0.12

Specimen362348NH

100
90
80

true stress, ksi

70
60
50
40
30

Avg. Yield Stress=59 ksi

20
10
0

0.02

0.04
0.06
0.08
true plastic strain
true plastic
strain, p

true stress,
true

0
0.001
0.002
0.007
0.012
0.017
0.027
0.037
0.047
0.057
0.067
0.077
0.087
0.097
0.107

ksi
34
50
56
64
68
72
78
82
86
88
91
92
94
96
97

483

0.1

0.12

Specimen362148H

100
90
80

true stress, ksi

70
60
50
40
30

Avg. Yield Stress=58.6 ksi

20
10
0

0.02

0.04
0.06
0.08
true plastic strain
true plastic
strain, p

true stress,
true

0
0.001
0.002
0.007
0.012
0.017
0.027
0.037
0.047
0.057
0.067
0.077
0.087
0.097
0.107

ksi
33.2
50.1
56.1
64.2
68.2
72.0
78.0
82.2
85.7
88.4
90.7
92.7
94.4
95.8
97.2

484

0.1

0.12

Specimen362248H

100
90
80

true stress, ksi

70
60
50
40
30

Avg. Yield Stress=59.7 ksi

20
10
0

0.02

0.04
0.06
0.08
true plastic strain
true plastic
strain, p

true stress,
true

0
0.001
0.002
0.007
0.012
0.017
0.027
0.037
0.047
0.057
0.067
0.077
0.087
0.097
0.107

ksi
37.3
51.5
56.7
64.2
68.2
72.0
78.1
82.4
85.8
88.5
90.8
92.7
94.4
95.9
97.2

485

0.1

0.12

Specimen362348H

100
90
80

true stress, ksi

70
60
50
40
30

Avg. Yield Stress=58.3 ksi

20
10
0

0.02

0.04
0.06
0.08
true plastic strain
true plastic
strain, p

true stress,
true

0
0.001
0.002
0.007
0.012
0.017
0.027
0.037
0.047
0.057
0.067
0.077
0.087
0.097
0.107

ksi
37.5
50.6
55.6
62.9
67.1
70.9
76.5
80.8
84.1
86.8
89.1
90.9
92.6
94.1
95.4

486

0.1

0.12

Specimen600124NH,600224NH,600324NH

100
90
80

true stress, ksi

70
60
50
40
30

Avg. Yield Stress=58.7 ksi

20
10
0

0.02

0.04
0.06
0.08
true plastic strain
true plastic
strain, p

true stress,
true

0
0.001
0.002
0.007
0.012
0.017
0.027
0.037
0.047
0.057
0.067
0.077
0.087
0.097
0.107

ksi
37.5
54.7
58.3
60.0
61.5
64.0
70.2
74.4
77.5
80.0
81.9
83.5
84.9
86.1
87.2

487

0.1

0.12

Specimen600124H

100
90
80

true stress, ksi

70
60
50
40
30

Avg. Yield Stress=61.9 ksi

20
10
0

0.02

0.04
0.06
0.08
true plastic strain
true plastic
strain, p

true stress,
true

0
0.001
0.002
0.007
0.012
0.017
0.027
0.037
0.047
0.057
0.067
0.077
0.087
0.097
0.107

ksi
35.0
58.9
61.4
62.3
62.9
63.6
69.4
74.4
77.8
80.5
82.6
84.4
85.9
87.2
88.4

488

0.1

0.12

Specimen600224H

100
90
80

true stress, ksi

70
60
50
40
30

Avg. Yield Stress=58.4 ksi

20
10
0

0.02

0.04
0.06
0.08
true plastic strain
true plastic
strain, p

true stress,
true

0
0.001
0.002
0.007
0.012
0.017
0.027
0.037
0.047
0.057
0.067
0.077
0.087
0.097
0.107

ksi
33.0
54.0
57.2
58.9
61.5
64.4
70.5
74.8
78.0
80.4
82.4
84.1
85.5
86.7
87.9

489

0.1

0.12

Specimen600324H

100
90
80

true stress, ksi

70
60
50
40
30

Avg. Yield Stress=60.1 ksi

20
10
0

0.02

0.04
0.06
0.08
true plastic strain
true plastic
strain, p

true stress,
true

0
0.001
0.002
0.007
0.012
0.017
0.027
0.037
0.047
0.057
0.067
0.077
0.087
0.097
0.107

ksi
34.9
57.5
60.1
60.8
61.8
63.6
69.8
74.1
77.6
79.9
82.0
83.7
85.1
86.5
87.6

490

0.1

0.12

Specimen600148NH

100
90
80

true stress, ksi

70
60
50
40
30

Avg. Yield Stress=60.1 ksi

20
10
0

0.02

0.04
0.06
0.08
true plastic strain
true plastic
strain, p

true stress,
true

0
0.001
0.002
0.007
0.012
0.017
0.027
0.037
0.047
0.057
0.067
0.077
0.087
0.097
0.107

ksi
44.4
58.9
60.3
61.0
62.0
64.8
70.6
75.0
78.3
80.8
82.8
84.5
86.0
87.3
88.5

491

0.1

0.12

Specimen600248NH

100
90
80

true stress, ksi

70
60
50
40
30

Avg. Yield Stress=63.4 ksi

20
10
0

0.02

0.04
0.06
0.08
true plastic strain
true plastic
strain, p

true stress,
true

0
0.001
0.002
0.007
0.012
0.017
0.027
0.037
0.047
0.057
0.067
0.077
0.087
0.097
0.107

ksi
41.9
62.1
63.1
63.6
64.0
64.5
69.8
75.0
78.5
81.2
83.5
85.3
87.0
88.4
89.6

492

0.1

0.12

Specimen600348NH

100
90
80

true stress, ksi

70
60
50
40
30

Avg. Yield Stress=61.2 ksi

20
10
0

0.02

0.04
0.06
0.08
true plastic strain
true plastic
strain, p

true stress,
true

0
0.001
0.002
0.007
0.012
0.017
0.027
0.037
0.047
0.057
0.067
0.077
0.087
0.097
0.107

ksi
36.0
55.6
60.0
61.7
62.0
62.5
68.5
73.4
76.9
79.4
81.6
83.3
84.9
86.2
87.4

493

0.1

0.12

Specimen600148H

100
90
80

true stress, ksi

70
60
50
40
30

Avg. Yield Stress=61.4 ksi

20
10
0

0.02

0.04
0.06
0.08
true plastic strain
true plastic
strain, p

true stress,
true

0
0.001
0.002
0.007
0.012
0.017
0.027
0.037
0.047
0.057
0.067
0.077
0.087
0.097
0.107

ksi
37.5
57.7
60.9
61.9
62.4
64.3
70.4
75.0
78.4
81.0
83.1
84.8
86.3
87.7
88.9

494

0.1

0.12

Specimen600248H

100
90
80

true stress, ksi

70
60
50
40
30

Avg. Yield Stress=62 ksi

20
10
0

0.02

0.04
0.06
0.08
true plastic strain
true plastic
strain, p

true stress,
true

0
0.001
0.002
0.007
0.012
0.017
0.027
0.037
0.047
0.057
0.067
0.077
0.087
0.097
0.107

ksi
37.4
57.6
61.6
62.6
63.3
64.1
69.4
74.2
77.8
80.6
82.9
84.8
86.4
87.8
89.1

495

0.1

0.12

Specimen600348H

100
90
80

true stress, ksi

70
60
50
40
30

Avg. Yield Stress=61.5 ksi

20
10
0

0.02

0.04
0.06
0.08
true plastic strain
true plastic
stra